Showing 51–100 of 172 results for author: Lau, R

Depth-aware Glass Surface Detection with Cross-modal Context Mining

Authors: Jiaying Lin, Yuen Hei Yeung, Rynson W. H. Lau

Abstract: Glass surfaces are becoming increasingly ubiquitous as modern buildings tend to use a lot of glass panels. This however poses substantial challenges on the operations of autonomous systems such as robots, self-driving cars and drones, as the glass panels can become transparent obstacles to the navigation.Existing works attempt to exploit various cues, including glass boundary context or reflection… ▽ More Glass surfaces are becoming increasingly ubiquitous as modern buildings tend to use a lot of glass panels. This however poses substantial challenges on the operations of autonomous systems such as robots, self-driving cars and drones, as the glass panels can become transparent obstacles to the navigation.Existing works attempt to exploit various cues, including glass boundary context or reflections, as a prior. However, they are all based on input RGB images.We observe that the transmission of 3D depth sensor light through glass surfaces often produces blank regions in the depth maps, which can offer additional insights to complement the RGB image features for glass surface detection. In this paper, we propose a novel framework for glass surface detection by incorporating RGB-D information, with two novel modules: (1) a cross-modal context mining (CCM) module to adaptively learn individual and mutual context features from RGB and depth information, and (2) a depth-missing aware attention (DAA) module to explicitly exploit spatial locations where missing depths occur to help detect the presence of glass surfaces. In addition, we propose a large-scale RGB-D glass surface detection dataset, called \textit{RGB-D GSD}, for RGB-D glass surface detection. Our dataset comprises 3,009 real-world RGB-D glass surface images with precise annotations. Extensive experimental results show that our proposed model outperforms state-of-the-art methods. △ Less

Submitted 22 June, 2022; originally announced June 2022.

SRGA J181414.6-225604: A new Galactic symbiotic X-ray binary outburst triggered by an intense mass loss episode of a heavily obscured Mira variable

Authors: Kishalay De, Ilya Mereminskiy, Roberto Soria, Charlie Conroy, Erin Kara, Shreya Anand, Michael C. B. Ashley, Martha L. Boyer, Deepto Chakrabarty, Brian Grefenstette, Matthew J. Hankins, Lynne A. Hillenbrand, Jacob E. Jencson, Viraj Karambelkar, Mansi M. Kasliwal, Ryan M. Lau, Alexander Lutovinov, Anna M. Moore, Mason Ng, Christos Panagiotou, Dheeraj R. Pasham, Andrey Semena, Robert Simcoe, Jamie Soon, Gokul P. Srinivasaragavan , et al. (2 additional authors not shown)

Abstract: We present the discovery and multi-wavelength characterization of SRGA J181414.6-225604, a Galactic hard X-ray transient discovered during the ongoing SRG/ART-XC sky survey. Using data from the Palomar Gattini-IR survey, we identify a spatially and temporally coincident variable infrared (IR) source, IRAS 18111-2257, and classify it as a very late-type (M7-M8), long period ($1502 \pm 24$ days) and… ▽ More We present the discovery and multi-wavelength characterization of SRGA J181414.6-225604, a Galactic hard X-ray transient discovered during the ongoing SRG/ART-XC sky survey. Using data from the Palomar Gattini-IR survey, we identify a spatially and temporally coincident variable infrared (IR) source, IRAS 18111-2257, and classify it as a very late-type (M7-M8), long period ($1502 \pm 24$ days) and luminous ($M_K\approx -9.9 \pm 0.2$) O-rich Mira donor star located at a distance of $\approx 14.6^{+2.9}_{-2.3}$ kpc. Combining multi-color photometric data over the last $\approx 25$ years, we show that the IR counterpart underwent a recent (starting $\approx 800$ days before the X-ray flare) enhanced mass loss (reaching $\approx 2.1 \times 10^{-5}$ M$_\odot$ yr$^{-1}$) episode resulting in an expanding dust shell obscuring the underlying star. Multi-epoch follow-up from Swift, NICER and NuSTAR reveal a $\approx 200$ day long X-ray outburst reaching a peak luminosity of $L_X \approx 2.5 \times 10^{36}$ erg s$^{-1}$, characterized by a heavily absorbed ($N_{\rm H} \approx 6\times 10^{22}$ cm$^{-2}$) X-ray spectrum consistent with an optically thick Comptonized plasma. The X-ray spectral and timing behavior suggest the presence of clumpy wind accretion together with a dense ionized nebula overabundant in silicate material surrounding the compact object. Together, we show that SRGA J181414.6-225604 is a new symbiotic X-ray binary in outburst, triggered by an intense dust formation episode of a highly evolved donor. Our results offer the first direct confirmation for the speculated connection between enhanced late-stage donor mass loss and active lifetimes of the symbiotic X-ray binaries. △ Less

Submitted 18 May, 2022; originally announced May 2022.

Comments: 32 pages, 15 figures, Accepted for publication in ApJ

Detection of a broad 8 $μ$m UIR feature in the mid-infrared spectrum of WR 125 observed with Subaru/COMICS

Authors: Izumi Endo, Ryan M. Lau, Itsuki Sakon, Takashi Onaka, Peredur M. Williams, Victor I. Shenavrin

Abstract: We present the detection of a broad 8 $μ$m feature in newly formed dust around the carbon-rich Wolf-Rayet (WC) binary WR 125 from N-band low-resolution (NL; R$\sim$250) spectroscopy between 7.3-13.6 $μ$m and N-band (11.7 $μ$m) and Q-band (18.8 $μ$m) imaging with Subaru/COMICS in 2019 October. WR 125 is a colliding wind binary (${\rm WC7+O9}$) that exhibited renewed dust formation starting in 2018,… ▽ More We present the detection of a broad 8 $μ$m feature in newly formed dust around the carbon-rich Wolf-Rayet (WC) binary WR 125 from N-band low-resolution (NL; R$\sim$250) spectroscopy between 7.3-13.6 $μ$m and N-band (11.7 $μ$m) and Q-band (18.8 $μ$m) imaging with Subaru/COMICS in 2019 October. WR 125 is a colliding wind binary (${\rm WC7+O9}$) that exhibited renewed dust formation starting in 2018, $\sim$28 years after its first dust formation episode had been observed. We also compare our infrared photometry with historical observations and revise the dust-formation period of WR 125 to 28.1 years. Archival infrared spectra of five dusty WC stars, WR 48a, WR 98a, WR 104, WR 112 and WR 118, obtained with ISO/SWS are reanalyzed and compared with the WR 125 spectrum to search for a similar feature. We analyze the dusty WC spectra using two different extinction curves to investigate the impact of interstellar extinction correction on the presence and/or properties of the 8 $μ$m feature. All of the dusty WC spectra dereddened with the two different extinction curves show a broad feature around 8 $μ$m (FWHM$\sim$1-2 $μ$m). We suggest that these 8 $μ$m features seen in the dusty WC spectra are related to the Class C unidentified infrared (UIR) features. △ Less

Submitted 8 May, 2022; v1 submitted 27 April, 2022; originally announced April 2022.

Comments: 17 pages, 7 figures, 4 tables, accepted to ApJ

Rethinking Video Salient Object Ranking

Authors: Jiaying Lin, Huankang Guan, Rynson W. H. Lau

Abstract: Salient Object Ranking (SOR) involves ranking the degree of saliency of multiple salient objects in an input image. Most recently, a method is proposed for ranking salient objects in an input video based on a predicted fixation map. It relies solely on the density of the fixations within the salient objects to infer their saliency ranks, which is incompatible with human perception of saliency rank… ▽ More Salient Object Ranking (SOR) involves ranking the degree of saliency of multiple salient objects in an input image. Most recently, a method is proposed for ranking salient objects in an input video based on a predicted fixation map. It relies solely on the density of the fixations within the salient objects to infer their saliency ranks, which is incompatible with human perception of saliency ranking. In this work, we propose to explicitly learn the spatial and temporal relations between different salient objects to produce the saliency ranks. To this end, we propose an end-to-end method for video salient object ranking (VSOR), with two novel modules: an intra-frame adaptive relation (IAR) module to learn the spatial relation among the salient objects in the same frame locally and globally, and an inter-frame dynamic relation (IDR) module to model the temporal relation of saliency across different frames. In addition, to address the limited video types (just sports and movies) and scene diversity in the existing VSOR dataset, we propose a new dataset that covers different video types and diverse scenes on a large scale. Experimental results demonstrate that our method outperforms state-of-the-art methods in relevant fields. We will make the source code and our proposed dataset available. △ Less

Submitted 31 March, 2022; originally announced March 2022.

Bi-directional Object-context Prioritization Learning for Saliency Ranking

Authors: Xin Tian, Ke Xu, Xin Yang, Lin Du, Baocai Yin, Rynson W. H. Lau

Abstract: The saliency ranking task is recently proposed to study the visual behavior that humans would typically shift their attention over different objects of a scene based on their degrees of saliency. Existing approaches focus on learning either object-object or object-scene relations. Such a strategy follows the idea of object-based attention in Psychology, but it tends to favor those objects with str… ▽ More The saliency ranking task is recently proposed to study the visual behavior that humans would typically shift their attention over different objects of a scene based on their degrees of saliency. Existing approaches focus on learning either object-object or object-scene relations. Such a strategy follows the idea of object-based attention in Psychology, but it tends to favor those objects with strong semantics (e.g., humans), resulting in unrealistic saliency ranking. We observe that spatial attention works concurrently with object-based attention in the human visual recognition system. During the recognition process, the human spatial attention mechanism would move, engage, and disengage from region to region (i.e., context to context). This inspires us to model the region-level interactions, in addition to the object-level reasoning, for saliency ranking. To this end, we propose a novel bi-directional method to unify spatial attention and object-based attention for saliency ranking. Our model includes two novel modules: (1) a selective object saliency (SOS) module that models objectbased attention via inferring the semantic representation of the salient object, and (2) an object-context-object relation (OCOR) module that allocates saliency ranks to objects by jointly modeling the object-context and context-object interactions of the salient objects. Extensive experiments show that our approach outperforms existing state-of-theart methods. Our code and pretrained model are available at https://github.com/GrassBro/OCOR. △ Less

Submitted 22 March, 2022; v1 submitted 17 March, 2022; originally announced March 2022.

Comments: Accepted to CVPR 2022

Filtration and breathability of nonwoven fabrics used in washable masks

Authors: Thomas W. Bement, Ania Mitros, Rebecca Lau, Timothy A. Sipkens, Jocelyn Songer, Heidi Alexander, Devon Ostrom, Hamed Nikookar, Steven. N. Rogak

Abstract: This study explores nonwoven and woven fabrics to improve upon the performance of the widespread all-cotton mask, and examines the effect of layering, machine washing and drying on their filtration and breathability for submicron and supermicron particles. Individual materials were evaluated for their quality factor, Q, which combines filtration efficiency and breathability. Filtration was tested… ▽ More This study explores nonwoven and woven fabrics to improve upon the performance of the widespread all-cotton mask, and examines the effect of layering, machine washing and drying on their filtration and breathability for submicron and supermicron particles. Individual materials were evaluated for their quality factor, Q, which combines filtration efficiency and breathability. Filtration was tested against particles 0.5 to 5 micron aerodynamic diameter. Nonwoven polyester and nonwoven polypropylene (craft fabrics, medical masks, and medical wraps) showed higher quality factors than woven materials (flannel cotton, Kona cotton, sateen cotton). Materials with meltblown nonwoven polypropylene filtered best, especially against submicron particles. Subsequently, we combined high performing fabrics into multi-layer sets, evaluating the sets quality factors before and after our washing protocol, which included machine washing, machine drying, and isopropanol soak. Sets incorporating meltblown nonwoven polypropylene designed for filtration (Filti and surgical mask) degraded significantly post-wash in the submicron range where they excelled prior to washing (Q = 57 and 79 at 1 micron, respectively, degraded to Q = 10 and 15 post-wash). Washing caused lesser quality degradation in sets incorporating spunbond non-woven polypropylene or medical wraps (Q = 12 to 24 pre-wash, Q = 8 to 10 post-wash). Post-wash quality factors are similar for all multi-layer sets in this study, and higher than Kona quilting cotton (Q = 6). Washed multi-layer sets filtered 12-42 percent of 0.5 micron, 27-76 percent of 1 micron, 58-96 percent of 2.8 micron, and 72-100 percent of 4.2 micron particles. The measured filtration and pressure drop of both the homogeneous and heterogeneous multi-layer fabric combinations agreed with the estimations from the layering model. △ Less

Submitted 7 February, 2022; originally announced February 2022.

Comments: 39 pages including supplemental information

A massive AGB donor in Scutum X-1: Identification of the first Mira variable in an X-ray binary

Authors: Kishalay De, Deepto Chakrabarty, Roberto Soria, Michael C. B. Ashley, Charlie Conroy, Matthew J. Hankins, Mansi M. Kasliwal, Ryan M. Lau, Anna M. Moore, Robert Simcoe, Jamie Soon, Tony Travouillon

Abstract: The symbiotic X-ray binary Sct X-1 was suggested as the first known neutron star accreting from a red supergiant companion. Although known for nearly 50 years, detailed characterization of the donor remains lacking, particularly due to the extremely high reddening towards the source ($A_V\gtrsim25$ mag). Here, we present i) improved localization of the counterpart using Gaia and Chandra observatio… ▽ More The symbiotic X-ray binary Sct X-1 was suggested as the first known neutron star accreting from a red supergiant companion. Although known for nearly 50 years, detailed characterization of the donor remains lacking, particularly due to the extremely high reddening towards the source ($A_V\gtrsim25$ mag). Here, we present i) improved localization of the counterpart using Gaia and Chandra observations, ii) the first broadband infrared spectrum ($\approx1-5\,μ$m; $R\approx 2000$) obtained with SpeX on the NASA Infrared Telescope Facility and iii) $J$-band light curve from the Palomar Gattini-IR survey. The infrared spectrum is characterized by i) deep water absorption features (H$_2$O index $\approx 40$%), ii) strong TiO, VO and CO features, and iii) weak/absent CN lines. We show that these features are inconsistent with known red supergiants, but suggest a M8-9 III type O-rich Mira donor star. We report the discovery of large amplitude ($ΔJ\approx3.5$ mag) periodic photometric variability suggesting a pulsation period of $621\pm36\,{\rm(systematic)}\pm8\,{\rm(statistical)}$ days, which we use to constrain the donor to be a relatively luminous Mira ($M_K=-8.6\pm0.3$ mag) at a distance of $3.6^{+0.8}_{-0.7}$ kpc. Comparing these characteristics to recent models, we find the donor to be consistent with a $\approx 3-5$ M$_\odot$ star at an age of $\approx 0.1-0.3$ Gyr. Together, we show that Sct X-1 was previously mis-classified as an evolved High Mass X-ray Binary; instead it is an intermediate mass system with the first confirmed Mira donor in an X-ray binary. We discuss the implications of Mira donors in symbiotic X-ray binaries, and highlight the potential of wide-field infrared time domain surveys and broadband infrared spectroscopy to unveil their demographics. △ Less

Submitted 2 March, 2022; v1 submitted 24 January, 2022; originally announced January 2022.

Comments: 10 pages, 4 figures. Accepted for publication in ApJ Letters

The Impact of Neutron Transfer Reactions on Heating and Cooling of Accreted Neutron Star Crusts

Authors: H. Schatz, Z. Meisel, E. F. Brown, S. S. Gupta, G. W. Hitt, W. R. Hix, R. Jain, R. Lau, P. Möller, W. -J. Ong, P. S. Shternin, Y. Xu, M. Wiescher

Abstract: Nuclear reactions heat and cool the crust of accreting neutron stars and need to be understood to interpret observations of X-ray bursts and of long-term cooling in transiently accreting systems. It was recently suggested that previously neglected neutron transfer reactions may play a significant role in the nuclear processes. We present results from full nuclear network calculations that now incl… ▽ More Nuclear reactions heat and cool the crust of accreting neutron stars and need to be understood to interpret observations of X-ray bursts and of long-term cooling in transiently accreting systems. It was recently suggested that previously neglected neutron transfer reactions may play a significant role in the nuclear processes. We present results from full nuclear network calculations that now include these reactions and determine their impact on crust composition, crust impurity, heating, and cooling. We find that a large number of neutron transfer reactions indeed occur and impact crust models. In particular, we identify a new type of reaction cycle that brings a pair of nuclei across the nuclear chart into equilibrium via alternating neutron capture and neutron release, interspersed with a neutron transfer. While neutron transfer reactions lead to changes in crust model predictions, and need to be considered in future studies, previous conclusions concerning heating, cooling, and compositional evolution are remarkably robust. △ Less

Submitted 21 December, 2021; originally announced December 2021.

Comments: 8 pages, 10 figures, accepted for Astrophysical Journal

Geometry-aware Two-scale PIFu Representation for Human Reconstruction

Authors: Zheng Dong, Ke Xu, Ziheng Duan, Hujun Bao, Weiwei Xu, Rynson W. H. Lau

Abstract: Although PIFu-based 3D human reconstruction methods are popular, the quality of recovered details is still unsatisfactory. In a sparse (e.g., 3 RGBD sensors) capture setting, the depth noise is typically amplified in the PIFu representation, resulting in flat facial surfaces and geometry-fallible bodies. In this paper, we propose a novel geometry-aware two-scale PIFu for 3D human reconstruction fr… ▽ More Although PIFu-based 3D human reconstruction methods are popular, the quality of recovered details is still unsatisfactory. In a sparse (e.g., 3 RGBD sensors) capture setting, the depth noise is typically amplified in the PIFu representation, resulting in flat facial surfaces and geometry-fallible bodies. In this paper, we propose a novel geometry-aware two-scale PIFu for 3D human reconstruction from sparse, noisy inputs. Our key idea is to exploit the complementary properties of depth denoising and 3D reconstruction, for learning a two-scale PIFu representation to reconstruct high-frequency facial details and consistent bodies separately. To this end, we first formulate depth denoising and 3D reconstruction as a multi-task learning problem. The depth denoising process enriches the local geometry information of the reconstruction features, while the reconstruction process enhances depth denoising with global topology information. We then propose to learn the two-scale PIFu representation using two MLPs based on the denoised depth and geometry-aware features. Extensive experiments demonstrate the effectiveness of our approach in reconstructing facial details and bodies of different poses and its superiority over state-of-the-art methods. △ Less

Submitted 27 September, 2022; v1 submitted 3 December, 2021; originally announced December 2021.

Comments: Accepted by NeurIPS 2022. 20 pages, 20 figures

Learning to Detect Instance-level Salient Objects Using Complementary Image Labels

Authors: Xin Tian, Ke Xu, Xin Yang, Baocai Yin, Rynson W. H. Lau

Abstract: Existing salient instance detection (SID) methods typically learn from pixel-level annotated datasets. In this paper, we present the first weakly-supervised approach to the SID problem. Although weak supervision has been considered in general saliency detection, it is mainly based on using class labels for object localization. However, it is non-trivial to use only class labels to learn instance-a… ▽ More Existing salient instance detection (SID) methods typically learn from pixel-level annotated datasets. In this paper, we present the first weakly-supervised approach to the SID problem. Although weak supervision has been considered in general saliency detection, it is mainly based on using class labels for object localization. However, it is non-trivial to use only class labels to learn instance-aware saliency information, as salient instances with high semantic affinities may not be easily separated by the labels. As the subitizing information provides an instant judgement on the number of salient items, it is naturally related to detecting salient instances and may help separate instances of the same class while grou** different parts of the same instance. Inspired by this observation, we propose to use class and subitizing labels as weak supervision for the SID problem. We propose a novel weakly-supervised network with three branches: a Saliency Detection Branch leveraging class consistency information to locate candidate objects; a Boundary Detection Branch exploiting class discrepancy information to delineate object boundaries; and a Centroid Detection Branch using subitizing information to detect salient instance centroids. This complementary information is then fused to produce a salient instance map. To facilitate the learning process, we further propose a progressive training scheme to reduce label noise and the corresponding noise learned by the model, via reciprocating the model with progressive salient instance prediction and model refreshing. Our extensive evaluations show that the proposed method plays favorably against carefully designed baseline methods adapted from related tasks. △ Less

Submitted 19 November, 2021; originally announced November 2021.

Comments: to appear IJCV. arXiv admin note: text overlap with arXiv:2009.13898

Second Timescale Photometry of the Very Fast Nova V1674 Her with Palomar Gattini-IR

Authors: Kylie Y. Hansen, Kishalay De, Michael C. B. Ashley, Mansi M. Kasliwal, Alexander Delacroix, Tim Greffe, David Hale, Matthew J. Hankins, Ryan Lau, Chengkui Li, Daniel McKenna, Anna M. Moore, Eran O. Ofek, Roger M. Smith, Jamie Soon, Roberto Soria, Gokul P. Srinivasaragavan, Tony Travouillon

Abstract: We report second-timescale infrared photometry of the nova V1674 Her using Palomar Gattini-IR. These observations constitute the first infrared and highest temporal resolution data (resolution of ~ 0.84 s) of the nova reported to date. PGIR observed in this fast readout mode for more than an hour on three nights between 3 and 6 days after discovery. We searched for periodic variability using a Lom… ▽ More We report second-timescale infrared photometry of the nova V1674 Her using Palomar Gattini-IR. These observations constitute the first infrared and highest temporal resolution data (resolution of ~ 0.84 s) of the nova reported to date. PGIR observed in this fast readout mode for more than an hour on three nights between 3 and 6 days after discovery. We searched for periodic variability using a Lomb-Scargle periodogram and did not detect anything down to a three sigma upper limit of 0.093 mag. This suggests that the periodic variability detected in the nova by Patterson et al. 2021 was lower by at least a factor of about 1.65 in the first week of the eruption. △ Less

Submitted 25 October, 2021; originally announced October 2021.

Comments: 3 pages, 1 figure

MODNet-V: Improving Portrait Video Matting via Background Restoration

Authors: Jiayu Sun, Zhanghan Ke, Lihe Zhang, Huchuan Lu, Rynson W. H. Lau

Abstract: To address the challenging portrait video matting problem more precisely, existing works typically apply some matting priors that require additional user efforts to obtain, such as annotated trimaps or background images. In this work, we observe that instead of asking the user to explicitly provide a background image, we may recover it from the input video itself. To this end, we first propose a n… ▽ More To address the challenging portrait video matting problem more precisely, existing works typically apply some matting priors that require additional user efforts to obtain, such as annotated trimaps or background images. In this work, we observe that instead of asking the user to explicitly provide a background image, we may recover it from the input video itself. To this end, we first propose a novel background restoration module (BRM) to recover the background image dynamically from the input video. BRM is extremely lightweight and can be easily integrated into existing matting models. By combining BRM with a recent image matting model, MODNet, we then present MODNet-V for portrait video matting. Benefited from the strong background prior provided by BRM, MODNet-V has only 1/3 of the parameters of MODNet but achieves comparable or even better performances. Our design allows MODNet-V to be trained in an end-to-end manner on a single NVIDIA 3090 GPU. Finally, we introduce a new patch refinement module (PRM) to adapt MODNet-V for high-resolution videos while kee** MODNet-V lightweight and fast. △ Less

Submitted 24 September, 2021; originally announced September 2021.

PGIR 20eid (SN2020qmp): A Type IIP Supernova at 15.6 Mpc discovered by the Palomar Gattini-IR survey

Authors: G. P. Srinivasaragavan, I. Sfaradi, J. Jencson, K. De, A. Horesh, M. M. Kasliwal, S. Tinyanont, M. Hankins, S. Schulze, M. C. B. Ashley, M. J. Graham, V. Karambelkar, R. Lau, A. A. Mahabal, A. M. Moore, E. O. Ofek, Y. Sharma, J. Sollerman, J. Soon, R. Soria, T. Travouillon, R. Walters

Abstract: We present a detailed analysis of SN 2020qmp, a nearby type IIP core-collapse supernova (CCSN), discovered by the Palomar Gattini-IR (PGIR) survey in the galaxy UGC07125. We illustrate how the multiwavelength study of this event helps our general understanding of stellar progenitors and circumstellar medium (CSM) interactions in CCSNe. We also highlight the importance of near-infrared (NIR) survey… ▽ More We present a detailed analysis of SN 2020qmp, a nearby type IIP core-collapse supernova (CCSN), discovered by the Palomar Gattini-IR (PGIR) survey in the galaxy UGC07125. We illustrate how the multiwavelength study of this event helps our general understanding of stellar progenitors and circumstellar medium (CSM) interactions in CCSNe. We also highlight the importance of near-infrared (NIR) surveys for early detections of SNe in dusty environments. SN 2020qmp displays characteristic hydrogen lines in its optical spectra, as well as a plateau in its optical LC, hallmarks of a type IIP SN. We do not detect linear polarization during the plateau phase, with a 3 sigma upper limit of 0.78%. Through hydrodynamical LC modeling and an analysis of its nebular spectra, we estimate a progenitor mass of around 11 solar masses, and an explosion energy of around 0.8e51 erg. We find that the spectral energy distribution cannot be explained by a simple CSM interaction model, assuming a constant shock velocity and steady mass-loss rate, and the excess X-ray luminosity compared with the synchrotron radio luminosity suggests deviations from equipartition. Finally, we demonstrate the advantages of NIR surveys over optical surveys for the detection of dust-obscured CCSNe in the local universe. Specifically, our simulations show that the Wide-Field Infrared Transient Explorer will detect about 14 more CCSNe out of 75 expected in its footprint within 40 Mpc, over five years than an optical survey equivalent to the Zwicky Transient Facility would detect. We have determined or constrained the main properties of SN 2020qmp and of its progenitor, highlighting the value of multiwavelength follow-up observations of nearby CCSNe. We have also shown that forthcoming NIR surveys will finally enable us to do a nearly complete census of CCSNe in the local universe. △ Less

Submitted 2 March, 2022; v1 submitted 5 September, 2021; originally announced September 2021.

Comments: 15 pages, 14 figures, Accepted in Astronomy & Astrophysics; Revised version after referee comments and language edits from journal

Journal ref: A&A 660, A138 (2022)

Scaled-Time-Attention Robust Edge Network

Authors: Richard Lau, Lihan Yao, Todd Huster, William Johnson, Stephen Arleth, Justin Wong, Devin Ridge, Michael Fletcher, William C. Headley

Abstract: This paper describes a systematic approach towards building a new family of neural networks based on a delay-loop version of a reservoir neural network. The resulting architecture, called Scaled-Time-Attention Robust Edge (STARE) network, exploits hyper dimensional space and non-multiply-and-add computation to achieve a simpler architecture, which has shallow layers, is simple to train, and is bet… ▽ More This paper describes a systematic approach towards building a new family of neural networks based on a delay-loop version of a reservoir neural network. The resulting architecture, called Scaled-Time-Attention Robust Edge (STARE) network, exploits hyper dimensional space and non-multiply-and-add computation to achieve a simpler architecture, which has shallow layers, is simple to train, and is better suited for Edge applications, such as Internet of Things (IoT), over traditional deep neural networks. STARE incorporates new AI concepts such as Attention and Context, and is best suited for temporal feature extraction and classification. We demonstrate that STARE is applicable to a variety of applications with improved performance and lower implementation complexity. In particular, we showed a novel way of applying a dual-loop configuration to detection and identification of drone vs bird in a counter Unmanned Air Systems (UAS) detection application by exploiting both spatial (video frame) and temporal (trajectory) information. We also demonstrated that the STARE performance approaches that of a State-of-the-Art deep neural network in classifying RF modulations, and outperforms Long Short-term Memory (LSTM) in a special case of Mackey Glass time series prediction. To demonstrate hardware efficiency, we designed and developed an FPGA implementation of the STARE algorithm to demonstrate its low-power and high-throughput operations. In addition, we illustrate an efficient structure for integrating a massively parallel implementation of the STARE algorithm for ASIC implementation. △ Less

Submitted 9 July, 2021; originally announced July 2021.

Comments: 20 pages, 22 figures, 9 tables, Darpa Distribution Statement A. Approved for public release. Distribution Unlimited

MSC Class: 68T05

Smart Scribbles for Image Mating

Authors: Xin Yang, Yu Qiao, Shaozhe Chen, Shengfeng He, Baocai Yin, Qiang Zhang, Xiaopeng Wei, Rynson W. H. Lau

Abstract: Image matting is an ill-posed problem that usually requires additional user input, such as trimaps or scribbles. Drawing a fne trimap requires a large amount of user effort, while using scribbles can hardly obtain satisfactory alpha mattes for non-professional users. Some recent deep learning-based matting networks rely on large-scale composite datasets for training to improve performance, resulti… ▽ More Image matting is an ill-posed problem that usually requires additional user input, such as trimaps or scribbles. Drawing a fne trimap requires a large amount of user effort, while using scribbles can hardly obtain satisfactory alpha mattes for non-professional users. Some recent deep learning-based matting networks rely on large-scale composite datasets for training to improve performance, resulting in the occasional appearance of obvious artifacts when processing natural images. In this article, we explore the intrinsic relationship between user input and alpha mattes and strike a balance between user effort and the quality of alpha mattes. In particular, we propose an interactive framework, referred to as smart scribbles, to guide users to draw few scribbles on the input images to produce high-quality alpha mattes. It frst infers the most informative regions of an image for drawing scribbles to indicate different categories (foreground, background, or unknown) and then spreads these scribbles (i.e., the category labels) to the rest of the image via our well-designed two-phase propagation. Both neighboring low-level afnities and high-level semantic features are considered during the propagation process. Our method can be optimized without large-scale matting datasets and exhibits more universality in real situations. Extensive experiments demonstrate that smart scribbles can produce more accurate alpha mattes with reduced additional input, compared to the state-of-the-art matting methods. △ Less

Submitted 31 March, 2021; originally announced March 2021.

Comments: ACM Trans. Multimedia Comput. Commun. Appl

Discovery of a 310-day Period from the Enshrouded Massive System NaSt1 (WR 122)

Authors: Ryan M. Lau, Samaporn Tinyanont, Matthew J. Hankins, Michael C. B. Ashley, Kishalay De, Alexei V. Filippenko, Lynne A. Hillenbrand, Mansi M. Kasliwal, Jon C. Mauerhan, Anthony F. J. Moffat, Anna M. Moore, Nathan Smith, Jamie Soon, Roberto Soria, Tony Travouillon, Karel A. van der Hucht, Peredur M. Williams, WeiKang Zheng

Abstract: We present optical and infrared (IR) light curves of NaSt1, also known as Wolf-Rayet (WR) 122, with observations from Palomar Gattini-IR (PGIR), the Zwicky Transient Facility (ZTF), the Katzman Automatic Imaging Telescope (KAIT), the Asteroid Terrestrial-impact Last Alert System (ATLAS), and the All-Sky Automated Survey for Supernovae (ASAS-SN). We identify a $P=309.7\pm0.7$ d photometric period f… ▽ More We present optical and infrared (IR) light curves of NaSt1, also known as Wolf-Rayet (WR) 122, with observations from Palomar Gattini-IR (PGIR), the Zwicky Transient Facility (ZTF), the Katzman Automatic Imaging Telescope (KAIT), the Asteroid Terrestrial-impact Last Alert System (ATLAS), and the All-Sky Automated Survey for Supernovae (ASAS-SN). We identify a $P=309.7\pm0.7$ d photometric period from the optical and IR light curves that reveal periodic, sinusoidal variability between 2014 July and 2021 July. We also present historical IR light curves taken between 1983 July and 1989 May, which show variability consistent with the period of the present-day light curves. In the past, NaSt1 was brighter in the $J$ band with larger variability amplitudes than the present-day PGIR values, suggesting that NaSt1 exhibits variability on longer ($\gtrsim$ decade) timescales. Sinusoidal fits to the recent optical and IR light curves show that the amplitude of NaSt1's variability differs at various wavelengths and also reveal significant phase offsets of $17.0\pm2.5$ d between the ZTF $r$ and PGIR $J$ light curves. We interpret the $310$ d photometric period from NaSt1 as the orbital period of an enshrouded massive binary. We suggest that the photometric variability of NaSt1 may arise from variations in the line-of-sight optical depth toward circumstellar optical/IR emitting regions throughout its orbit due to colliding-wind dust formation. We speculate that past mass transfer in NaSt1 may have been triggered by Roche-lobe overflow (RLOF) during an eruptive phase of an Ofpe/WN9 star. Lastly, we argue that NaSt1 is no longer undergoing RLOF mass transfer. △ Less

Submitted 29 August, 2021; v1 submitted 15 March, 2021; originally announced March 2021.

Comments: 20 Pages, 7 Figures, 5 Tables, 1 Table Attachment, Accepted to ApJ

AT 2019qyl in NGC 300: Internal Collisions in the Early Outflow from a Very Fast Nova in a Symbiotic Binary

Authors: Jacob E. Jencson, Jennifer E. Andrews, Howard E. Bond, Viraj Karambelkar, David J. Sand, Schuyler D. van Dyk, Nadejda Blagorodnova, Martha L. Boyer, Mansi M. Kasliwal, Ryan M. Lau, Shazrene Mohamed, Robert Williams, Patricia A. Whitelock, Rachael C. Amaro, K. Azalee Bostroem, Yize Dong, Michael J. Lundquist, Stefano Valenti, Samuel D. Wyatt, Jamie Burke, Kishalay De, Saurabh W. Jha, Joel Johansson, César Rojas-Bravo, David A. Coulter , et al. (17 additional authors not shown)

Abstract: Nova eruptions, thermonuclear explosions on the surfaces of white dwarfs (WDs), are now recognized to be among the most common shock-powered astrophysical transients. We present the early discovery and rapid ultraviolet (UV), optical, and infrared (IR) temporal development of AT 2019qyl, a recent nova in the nearby Sculptor Group galaxy NGC 300. The light curve shows a rapid rise lasting… ▽ More Nova eruptions, thermonuclear explosions on the surfaces of white dwarfs (WDs), are now recognized to be among the most common shock-powered astrophysical transients. We present the early discovery and rapid ultraviolet (UV), optical, and infrared (IR) temporal development of AT 2019qyl, a recent nova in the nearby Sculptor Group galaxy NGC 300. The light curve shows a rapid rise lasting $\lesssim 1$ day, reaching a peak absolute magnitude of $M_V = -9.2$ mag, and a very fast decline, fading by 2 mag over 3.5 days. A steep dropoff in the light curves after 71 days and the rapid decline timescale suggest a low-mass ejection from a massive WD with $M_{\rm WD} \gtrsim 1.2~M_{\odot}$. We present an unprecedented view of the early spectroscopic evolution of such an event. Three spectra prior to the peak reveal a complex, multicomponent outflow giving rise to internal collisions and shocks in the ejecta of an He/N-class nova. We identify a coincident IR-variable counterpart in the extensive preeruption coverage of the transient location and infer the presence of a symbiotic progenitor system with an O-rich asymptotic-giant-branch donor star, as well as evidence for an earlier UV-bright outburst in 2014. We suggest that AT 2019qyl is analogous to the subset of Galactic recurrent novae with red-giant companions such as RS Oph and other embedded nova systems like V407 Cyg. Our observations provide new evidence that internal shocks between multiple, distinct outflow components likely contribute to the generation of the shock-powered emission from such systems. △ Less

Submitted 5 September, 2021; v1 submitted 22 February, 2021; originally announced February 2021.

Comments: 32 pages, 10 figures, accepted in ApJ

The luminous red nova AT 2018bwo in NGC 45 and its binary yellow supergiant progenitor

Authors: Nadejda Blagorodnova, Jakub Klencki, Ondrej Pejcha, Paul M. Vreeswijk, Howard E. Bond, Kevin B. Burdge, Kishalay De, Christoffer Fremling, Robert D. Gehrz, Jacob E. Jencson, Mansi M. Kasliwal, Thomas Kupfer, Ryan M. Lau, Frank J. Masci, R. Michael Rich

Abstract: Luminous Red Novae (LRNe) are astrophysical transients associated with the partial ejection of a binary system's common envelope (CE) shortly before its merger. Here we present the results of our photometric and spectroscopic follow-up campaign of AT2018bwo (DLT18x), a LRN discovered in NGC45, and investigate its progenitor system using binary stellar-evolution models. The transient reached a peak… ▽ More Luminous Red Novae (LRNe) are astrophysical transients associated with the partial ejection of a binary system's common envelope (CE) shortly before its merger. Here we present the results of our photometric and spectroscopic follow-up campaign of AT2018bwo (DLT18x), a LRN discovered in NGC45, and investigate its progenitor system using binary stellar-evolution models. The transient reached a peak magnitude of $M_r=-10.97\pm0.11$ and maintained this brightness during its optical plateau of $t_p = 41\pm5$days. During this phase, it showed a rather stable photospheric temperature of ~3300K and a luminosity of ~$10^{40}$erg/s. The photosphere of AT2018bwo at early times appeared larger and cooler than other similar LRNe, likely due to an extended mass-loss episode before the merger. Towards the end of the plateau, optical spectra showed a reddened continuum with strong molecular absorption bands. The reprocessed emission by the cooling dust was also detected in the mid-infrared bands ~1.5 years after the outburst. Archival Spitzer and Hubble Space Telescope data taken 10-14 years before the transient event suggest a progenitor star with $T_{prog}\sim 6500$K, $R_{prog}\sim 100R_{\odot}$ and $L_{prog}\sim 2\times10^4L_{\odot}$, and an upper limit for optically thin warm (1000 K) dust mass of $M_d<10^{-6}M_{\odot}$. Using stellar binary-evolution models, we determined the properties of binary systems consistent with the progenitor parameter space. For AT2018bwo, we infer a primary mass of 12-16 $M_{\odot}$, which is 9-45% larger than the ~11$M_{\odot}$ obtained using single-star evolution models. The system, consistent with a yellow-supergiant primary, was likely in a stable mass-transfer regime with -2.4<log ($\dot{M}/M_{\odot}$/yr)<-1.2 a decade before the main instability occurred. During the dynamical merger, the system would have ejected 0.15-0.5$M_{\odot}$ with a velocity of ~500 km/s. △ Less

Submitted 27 January, 2022; v1 submitted 10 February, 2021; originally announced February 2021.

Comments: 21 pages, 16 figures, accepted version for A&A

Journal ref: A&A 653, A134 (2021)

Automatic Comic Generation with Stylistic Multi-page Layouts and Emotion-driven Text Balloon Generation

Authors: Xin Yang, Zongliang Ma, Letian Yu, Ying Cao, Baocai Yin, Xiaopeng Wei, Qiang Zhang, Rynson W. H. Lau

Abstract: In this paper, we propose a fully automatic system for generating comic books from videos without any human intervention. Given an input video along with its subtitles, our approach first extracts informative keyframes by analyzing the subtitles, and stylizes keyframes into comic-style images. Then, we propose a novel automatic multi-page layout framework, which can allocate the images across mult… ▽ More In this paper, we propose a fully automatic system for generating comic books from videos without any human intervention. Given an input video along with its subtitles, our approach first extracts informative keyframes by analyzing the subtitles, and stylizes keyframes into comic-style images. Then, we propose a novel automatic multi-page layout framework, which can allocate the images across multiple pages and synthesize visually interesting layouts based on the rich semantics of the images (e.g., importance and inter-image relation). Finally, as opposed to using the same type of balloon as in previous works, we propose an emotion-aware balloon generation method to create different types of word balloons by analyzing the emotion of subtitles and audios. Our method is able to vary balloon shapes and word sizes in balloons in response to different emotions, leading to more enriched reading experience. Once the balloons are generated, they are placed adjacent to their corresponding speakers via speaker detection. Our results show that our method, without requiring any user inputs, can generate high-quality comic pages with visually rich layouts and balloons. Our user studies also demonstrate that users prefer our generated results over those by state-of-the-art comic generation systems. △ Less

Submitted 26 January, 2021; originally announced January 2021.

Outbursting Young Stellar Object PGIR 20dci in the Perseus Arm

Authors: Lynne A. Hillenbrand, Kishalay De, Matthew Hankins, Mansi M. Kasliwal, Luisa M. Rebull, Ryan M. Lau, Roc M. Cutri, Michael C. B. Ashley, Viraj R. Karambelkar, Anna M. Moore, T. Travouillon, A. K. Mainzer

Abstract: We report the discovery of a likely outbursting Class I young stellar object, associated with the star-forming region NGC 281-W (distance $\sim 2.8$ kpc). The source is currently seen only at infrared wavelengths, appearing in both the Palomar Gattini InfraRed ($1.2~μ$m) and the Near Earth Object Widefield Infrared Survey Explorer ($3.4$ and $4.6~μ$m) photometric time-domain surveys. Recent near-i… ▽ More We report the discovery of a likely outbursting Class I young stellar object, associated with the star-forming region NGC 281-W (distance $\sim 2.8$ kpc). The source is currently seen only at infrared wavelengths, appearing in both the Palomar Gattini InfraRed ($1.2~μ$m) and the Near Earth Object Widefield Infrared Survey Explorer ($3.4$ and $4.6~μ$m) photometric time-domain surveys. Recent near-infrared imaging reveals a new, extended scattered light nebula. Recent near-infrared spectroscopy confirms the similarity of PGIR 20dci to FU Ori type sources, based on strong molecular absorption in CO, H$_2$O, and OH, weak absorption in several atomic lines, and a warm wind/outflow as indicated by a P Cygni profile in the HeI 10830 A line. This is a rare case of an FU Ori star with a well-measured long term photometric rise before a sharper outburst, and the second instance of an FU Ori star with a documented two-step brightening in the mid-infrared. △ Less

Submitted 9 February, 2021; v1 submitted 11 January, 2021; originally announced January 2021.

Comments: accepted to AAS Journals on 2021, February 5

A population of heavily reddened, optically missed novae from Palomar Gattini-IR: Constraints on the Galactic nova rate

Authors: Kishalay De, Mansi M. Kasliwal, Matthew J. Hankins, Jennifer L. Sokoloski, Scott M. Adams, Michael C. B. Ashley, Aliya-Nur Babul, Ashot Bagdasaryan, Alexandre Delacroix, Richard Dekany, Timothee Greffe, David Hale, Jacob E. Jencson, Viraj R. Karambelkar, Ryan M. Lau, Ashish Mahabal, Daniel McKenna, Anna M. Moore, Eran O. Ofek, Manasi Sharma, Roger M. Smith, Jamie Soon, Roberto Soria, Gokul Srinivasaragavan, Samaporn Tinyanont , et al. (3 additional authors not shown)

Abstract: The nova rate in the Milky Way remains largely uncertain, despite its vital importance in constraining models of Galactic chemical evolution as well as understanding progenitor channels for Type Ia supernovae. The rate has been previously estimated in the range of $\approx10-300$ yr$^{-1}$, either based on extrapolations from a handful of very bright optical novae or the nova rates in nearby galax… ▽ More The nova rate in the Milky Way remains largely uncertain, despite its vital importance in constraining models of Galactic chemical evolution as well as understanding progenitor channels for Type Ia supernovae. The rate has been previously estimated in the range of $\approx10-300$ yr$^{-1}$, either based on extrapolations from a handful of very bright optical novae or the nova rates in nearby galaxies; both methods are subject to debatable assumptions. The total discovery rate of optical novae remains much smaller ($\approx5-10$ yr$^{-1}$) than these estimates, even with the advent of all-sky optical time domain surveys. Here, we present a systematic sample of 12 spectroscopically confirmed Galactic novae detected in the first 17 months of Palomar Gattini-IR (PGIR), a wide-field near-infrared time domain survey. Operating in $J$-band ($\approx1.2$ $μ$m) that is relatively immune to dust extinction, the extinction distribution of the PGIR sample is highly skewed to large extinction values ($> 50$% of events obscured by $A_V\gtrsim5$ mag). Using recent estimates for the distribution of mass and dust in the Galaxy, we show that the observed extinction distribution of the PGIR sample is commensurate with that expected from dust models. The PGIR extinction distribution is inconsistent with that reported in previous optical searches (null hypothesis probability $< 0.01$%), suggesting that a large population of highly obscured novae have been systematically missed in previous optical searches. We perform the first quantitative simulation of a $3π$ time domain survey to estimate the Galactic nova rate using PGIR, and derive a rate of $\approx 46.0^{+12.5}_{-12.4}$ yr$^{-1}$. Our results suggest that all-sky near-infrared time-domain surveys are well poised to uncover the Galactic nova population. △ Less

Submitted 24 January, 2021; v1 submitted 11 January, 2021; originally announced January 2021.

Comments: 25 pages, 12 figures. Submitted to ApJ

Weakly-Supervised Saliency Detection via Salient Object Subitizing

Authors: Xiaoyang Zheng, Xin Tan, Jie Zhou, Lizhuang Ma, Rynson W. H. Lau

Abstract: Salient object detection aims at detecting the most visually distinct objects and producing the corresponding masks. As the cost of pixel-level annotations is high, image tags are usually used as weak supervisions. However, an image tag can only be used to annotate one class of objects. In this paper, we introduce saliency subitizing as the weak supervision since it is class-agnostic. This allows… ▽ More Salient object detection aims at detecting the most visually distinct objects and producing the corresponding masks. As the cost of pixel-level annotations is high, image tags are usually used as weak supervisions. However, an image tag can only be used to annotate one class of objects. In this paper, we introduce saliency subitizing as the weak supervision since it is class-agnostic. This allows the supervision to be aligned with the property of saliency detection, where the salient objects of an image could be from more than one class. To this end, we propose a model with two modules, Saliency Subitizing Module (SSM) and Saliency Updating Module (SUM). While SSM learns to generate the initial saliency masks using the subitizing information, without the need for any unsupervised methods or some random seeds, SUM helps iteratively refine the generated saliency masks. We conduct extensive experiments on five benchmark datasets. The experimental results show that our method outperforms other weakly-supervised methods and even performs comparably to some fully-supervised methods. △ Less

Submitted 4 January, 2021; originally announced January 2021.

Comments: This paper is accepted to IEEE Trans. on Circuits and Systems for Video Technology (TCSVT)

Census of R Coronae Borealis stars I: Infrared light curves from Palomar Gattini IR

Authors: Viraj R. Karambelkar, Mansi M. Kasliwal, Patrick Tisserand, Kishalay De, Shreya Anand, Michael C. B. Ashley, Alex Delacroix, Matthew Hankins, Jacob E. Jencson, Ryan M. Lau, Dan McKenna, Anna Moore, Eran O. Ofek, Roger M. Smith, Roberto Soria, Jamie Soon, Samaporn Tinyanont, Tony Travouillon, Yuhan Yao

Abstract: We are undertaking the first systematic infrared (IR) census of R Coronae Borealis (RCB) stars in the Milky Way, beginning with IR light curves from the Palomar Gattini IR (PGIR) survey. PGIR is a 30 cm $J$-band telescope with a 25 deg$^{2}$ camera that is surveying 18000 deg$^{2}$ of the northern sky ($δ>-28^{o}$) at a cadence of 2 days. We present PGIR light curves for 922 RCB candidates selecte… ▽ More We are undertaking the first systematic infrared (IR) census of R Coronae Borealis (RCB) stars in the Milky Way, beginning with IR light curves from the Palomar Gattini IR (PGIR) survey. PGIR is a 30 cm $J$-band telescope with a 25 deg$^{2}$ camera that is surveying 18000 deg$^{2}$ of the northern sky ($δ>-28^{o}$) at a cadence of 2 days. We present PGIR light curves for 922 RCB candidates selected from a mid-IR color-based catalog (Tisserand et al. 2020). Of these 922, 149 are promising RCB candidates as they show pulsations or declines similar to RCB stars. Majority of the candidates that are not RCB stars are either long period variables (LPVs) or RV-Tauri stars. We identify IR color-based criteria to better distinguish between RCB stars and LPVs. As part of a pilot spectroscopic run, we obtained NIR spectra for 26 out of the 149 promising candidates and spectroscopically confirm 11 new RCB stars. We detect strong He I $λ10830$ features in spectra of all RCB stars, likely originating within high velocity (200-400 km-s$^{-1}$) winds in their atmospheres. 9 of these RCB stars show $^{12}$C$^{16}$O and $^{12}$C$^{18}$O molecular absorption features, suggesting that they are formed through a white dwarf merger. We detect quasi-periodic pulsations in the light curves of 5 RCB stars. The periods range between 30-125 days and likely originate from the strange-mode instability in these stars. Our pilot run results motivate a dedicated IR spectroscopic campaign to classify all RCB candidates. △ Less

Submitted 21 December, 2020; originally announced December 2020.

Comments: 20 pages, 14 figures, Submitted to ApJ

Location-aware Single Image Reflection Removal

Authors: Zheng Dong, Ke Xu, Yin Yang, Hujun Bao, Weiwei Xu, Rynson W. H. Lau

Abstract: This paper proposes a novel location-aware deep-learning-based single image reflection removal method. Our network has a reflection detection module to regress a probabilistic reflection confidence map, taking multi-scale Laplacian features as inputs. This probabilistic map tells if a region is reflection-dominated or transmission-dominated, and it is used as a cue for the network to control the f… ▽ More This paper proposes a novel location-aware deep-learning-based single image reflection removal method. Our network has a reflection detection module to regress a probabilistic reflection confidence map, taking multi-scale Laplacian features as inputs. This probabilistic map tells if a region is reflection-dominated or transmission-dominated, and it is used as a cue for the network to control the feature flow when predicting the reflection and transmission layers. We design our network as a recurrent network to progressively refine reflection removal results at each iteration. The novelty is that we leverage Laplacian kernel parameters to emphasize the boundaries of strong reflections. It is beneficial to strong reflection detection and substantially improves the quality of reflection removal results. Extensive experiments verify the superior performance of the proposed method over state-of-the-art approaches. Our code and the pre-trained model can be found at https://github.com/zdlarr/Location-aware-SIRR. △ Less

Submitted 19 August, 2021; v1 submitted 13 December, 2020; originally announced December 2020.

Comments: 10 pages, 10 figures, 3 tables

MODNet: Real-Time Trimap-Free Portrait Matting via Objective Decomposition

Authors: Zhanghan Ke, Jiayu Sun, Kaican Li, Qiong Yan, Rynson W. H. Lau

Abstract: Existing portrait matting methods either require auxiliary inputs that are costly to obtain or involve multiple stages that are computationally expensive, making them less suitable for real-time applications. In this work, we present a light-weight matting objective decomposition network (MODNet) for portrait matting in real-time with a single input image. The key idea behind our efficient design… ▽ More Existing portrait matting methods either require auxiliary inputs that are costly to obtain or involve multiple stages that are computationally expensive, making them less suitable for real-time applications. In this work, we present a light-weight matting objective decomposition network (MODNet) for portrait matting in real-time with a single input image. The key idea behind our efficient design is by optimizing a series of sub-objectives simultaneously via explicit constraints. In addition, MODNet includes two novel techniques for improving model efficiency and robustness. First, an Efficient Atrous Spatial Pyramid Pooling (e-ASPP) module is introduced to fuse multi-scale features for semantic estimation. Second, a self-supervised sub-objectives consistency (SOC) strategy is proposed to adapt MODNet to real-world data to address the domain shift problem common to trimap-free methods. MODNet is easy to be trained in an end-to-end manner. It is much faster than contemporaneous methods and runs at 67 frames per second on a 1080Ti GPU. Experiments show that MODNet outperforms prior trimap-free methods by a large margin on both Adobe Matting Dataset and a carefully designed photographic portrait matting (PPM-100) benchmark proposed by us. Further, MODNet achieves remarkable results on daily photos and videos. Our code and models are available at https://github.com/ZHKKKe/MODNet, and the PPM-100 benchmark is released at https://github.com/ZHKKKe/PPM. △ Less

Submitted 18 March, 2022; v1 submitted 24 November, 2020; originally announced November 2020.

Revealing Efficient Dust-Formation at Low Metallicity in Extragalactic Carbon-Rich Wolf-Rayet Binaries

Authors: Ryan M. Lau, Matthew J. Hankins, Mansi M. Kasliwal, Howard E. Bond, Kishalay De, Jacob E. Jencson, Anthony F. J. Moffat, Nathan Smith, Peredur M. Williams

Abstract: We present Spitzer/IRAC observations of dust formation from six extragalactic carbon-rich Wolf-Rayet (WC) binary candidates in low-metallicity (Z $\lesssim0.65$ Z$_\odot$) environments using multi-epoch mid-infrared (IR) imaging data from the SPitzer InfraRed Intensive Transients Survey (SPIRITS). Optical follow-up spectroscopy of SPIRITS~16ln, 19q, 16df, 18hb, and 14apu reveals emission features… ▽ More We present Spitzer/IRAC observations of dust formation from six extragalactic carbon-rich Wolf-Rayet (WC) binary candidates in low-metallicity (Z $\lesssim0.65$ Z$_\odot$) environments using multi-epoch mid-infrared (IR) imaging data from the SPitzer InfraRed Intensive Transients Survey (SPIRITS). Optical follow-up spectroscopy of SPIRITS~16ln, 19q, 16df, 18hb, and 14apu reveals emission features from C IV $\lambda5801\text{-}12$~and/or the C III-IV $\lambda4650$ He II $\lambda4686$~blend that are consistent with early-type WC stars. We identify SPIRITS~16ln as the variable mid-IR counterpart of the recently discovered colliding-wind WC4+O binary candidate, N604-WRXc, located in the sub-solar metallicity NGC 604 H II~region in M33. We interpret the mid-IR variability from SPIRITS~16ln as a dust-formation episode in an eccentric colliding-wind WC binary. SPIRITS~19q, 16df, 14apu, and 18hb exhibit absolute [3.6] magnitudes exceeding one of most IR-luminous dust-forming WC systems known, WR~104 (M$_\mathrm{[3.6]}\lesssim-12.3$). An analysis of dust formation in the mid-IR outburst from SPIRITS~19q reveals a high dust production rate of $\dot{M}_d\gtrsim2\times10^{-6}$ M$_\odot$ yr$^{-1}$, which may therefore exceed that of the most efficient dust-forming WC systems known. We demonstrate that efficient dust-formation is feasible from early-type WC binaries in the theoretical framework of colliding-wind binary dust formation if the systems host an O-type companion with a high mass-loss rate ($\dot{M}\gtrsim1.6\times10^{-6}$ M$_\odot$ yr$^{-1}$). This efficient dust-formation from early-type WC binaries highlights their potential role as significant sources of dust in low-metallicity environments. △ Less

Submitted 6 January, 2021; v1 submitted 19 November, 2020; originally announced November 2020.

Comments: 28 pages, 10 Figures, and 5 Tables. Accepted to ApJ on 2021 Jan 4 (Replacement of version submitted to arXiv on 2020 Nov 19)

Weakly-supervised Salient Instance Detection

Authors: Xin Tian, Ke Xu, Xin Yang, Baocai Yin, Rynson W. H. Lau

Abstract: Existing salient instance detection (SID) methods typically learn from pixel-level annotated datasets. In this paper, we present the first weakly-supervised approach to the SID problem. Although weak supervision has been considered in general saliency detection, it is mainly based on using class labels for object localization. However, it is non-trivial to use only class labels to learn instance-a… ▽ More Existing salient instance detection (SID) methods typically learn from pixel-level annotated datasets. In this paper, we present the first weakly-supervised approach to the SID problem. Although weak supervision has been considered in general saliency detection, it is mainly based on using class labels for object localization. However, it is non-trivial to use only class labels to learn instance-aware saliency information, as salient instances with high semantic affinities may not be easily separated by the labels. We note that subitizing information provides an instant judgement on the number of salient items, which naturally relates to detecting salient instances and may help separate instances of the same class while grou** different parts of the same instance. Inspired by this insight, we propose to use class and subitizing labels as weak supervision for the SID problem. We propose a novel weakly-supervised network with three branches: a Saliency Detection Branch leveraging class consistency information to locate candidate objects; a Boundary Detection Branch exploiting class discrepancy information to delineate object boundaries; and a Centroid Detection Branch using subitizing information to detect salient instance centroids. This complementary information is further fused to produce salient instance maps. We conduct extensive experiments to demonstrate that the proposed method plays favorably against carefully designed baseline methods adapted from related tasks. △ Less

Submitted 29 September, 2020; originally announced September 2020.

Comments: BMVC 2020, best student paper runner-up

The extreme colliding-wind system Apep: resolved imagery of the central binary and dust plume in the infrared

Authors: Y. Han, P. G. Tuthill, R. M. Lau, A. Soulain, J. R. Callingham, P. M. Williams, P. A. Crowther, B. J. S. Pope, B. Marcote

Abstract: The recent discovery of a spectacular dust plume in the system 2XMM J160050.7-514245 (referred to as "Apep") suggested a physical origin in a colliding-wind binary by way of the "Pinwheel" mechanism. Observational data pointed to a hierarchical triple-star system, however several extreme and unexpected physical properties seem to defy the established physics of such objects. Most notably, a stark… ▽ More The recent discovery of a spectacular dust plume in the system 2XMM J160050.7-514245 (referred to as "Apep") suggested a physical origin in a colliding-wind binary by way of the "Pinwheel" mechanism. Observational data pointed to a hierarchical triple-star system, however several extreme and unexpected physical properties seem to defy the established physics of such objects. Most notably, a stark discrepancy was found in the observed outflow speed of the gas as measured spectroscopically in the line-of-sight direction compared to the proper motion expansion of the dust in the sky plane. This enigmatic behaviour arises at the wind base within the central Wolf-Rayet binary: a system that has so far remained spatially unresolved. Here we present an updated proper motion study deriving the expansion speed of Apep's dust plume over a two-year baseline that is four times slower than the spectroscopic wind speed, confirming and strengthening the previous finding. We also present the results from high-angular-resolution near-infrared imaging studies of the heart of the system, revealing a close binary with properties matching a Wolf-Rayet colliding-wind system. Based on these new observational constraints, an improved geometric model is presented yielding a close match to the data, constraining the orbital parameters of the Wolf-Rayet binary and lending further support to the anisotropic wind model. △ Less

Submitted 13 August, 2020; originally announced August 2020.

Comments: This article has been accepted for publication in the Monthly Notices of the Royal Astronomical Society. 17 pages, 9 figures, 6 tables

Guided Collaborative Training for Pixel-wise Semi-Supervised Learning

Authors: Zhanghan Ke, Di Qiu, Kaican Li, Qiong Yan, Rynson W. H. Lau

Abstract: We investigate the generalization of semi-supervised learning (SSL) to diverse pixel-wise tasks. Although SSL methods have achieved impressive results in image classification, the performances of applying them to pixel-wise tasks are unsatisfactory due to their need for dense outputs. In addition, existing pixel-wise SSL approaches are only suitable for certain tasks as they usually require to use… ▽ More We investigate the generalization of semi-supervised learning (SSL) to diverse pixel-wise tasks. Although SSL methods have achieved impressive results in image classification, the performances of applying them to pixel-wise tasks are unsatisfactory due to their need for dense outputs. In addition, existing pixel-wise SSL approaches are only suitable for certain tasks as they usually require to use task-specific properties. In this paper, we present a new SSL framework, named Guided Collaborative Training (GCT), for pixel-wise tasks, with two main technical contributions. First, GCT addresses the issues caused by the dense outputs through a novel flaw detector. Second, the modules in GCT learn from unlabeled data collaboratively through two newly proposed constraints that are independent of task-specific properties. As a result, GCT can be applied to a wide range of pixel-wise tasks without structural adaptation. Our extensive experiments on four challenging vision tasks, including semantic segmentation, real image denoising, portrait image matting, and night image enhancement, show that GCT outperforms state-of-the-art SSL methods by a large margin. Our code available at: https://github.com/ZHKKKe/PixelSSL. △ Less

Submitted 12 August, 2020; originally announced August 2020.

Comments: 16th European Conference on Computer Vision (ECCV 2020)

Resolving Decades of Periodic Spirals from the Wolf-Rayet Dust Factory WR 112

Authors: Ryan M. Lau, Matthew J. Hankins, Yinuo Han, Izumi Endo, Anthony F. J. Moffat, Michael E. Ressler, Itsuki Sakon, Joel Sanchez-Bermudez, Anthony Soulain, Ian R. Stevens, Peter G. Tuthill, Peredur M. Williams

Abstract: WR 112 is a dust-forming carbon-rich Wolf-Rayet (WC) binary with a dusty circumstellar nebula that exhibits a complex asymmetric morphology, which traces the orbital motion and dust formation in the colliding winds of the central binary. Unraveling the complicated circumstellar dust emission around WR 112 therefore provides an opportunity to understand the dust formation process in colliding-wind… ▽ More WR 112 is a dust-forming carbon-rich Wolf-Rayet (WC) binary with a dusty circumstellar nebula that exhibits a complex asymmetric morphology, which traces the orbital motion and dust formation in the colliding winds of the central binary. Unraveling the complicated circumstellar dust emission around WR 112 therefore provides an opportunity to understand the dust formation process in colliding-wind WC binaries. In this work, we present a multi-epoch analysis of the circumstellar dust around WR 112 using seven high spatial resolution (FWHM $\sim0.3-0.4''$) N-band ($λ\sim12$ $μ$m) imaging observations spanning almost 20 years and includes newly obtained images from Subaru/COMICS in Oct 2019. In contrast to previous interpretations of a face-on spiral morphology, we observe clear evidence of proper motion of the circumstellar dust around WR 112 consistent with a nearly edge-on spiral with a $θ_s=55^\circ$ half-opening angle and a $\sim20$-yr period. The revised near edge-on geometry of WR 112 reconciles previous observations of highly variable non-thermal radio emission that was inconsistent with a face-on geometry. We estimate a revised distance to WR 112 of $d = 3.39^{+0.89}_{-0.84}$ kpc based on the observed dust expansion rate and a spectroscopically derived WC terminal wind velocity of $v_\infty= 1230\pm260$ km s$^{-1}$. With the newly derived WR 112 parameters we fit optically-thin dust spectral energy distribution models and determine a dust production rate of $\dot{M}_d=2.7^{+1.0}_{-1.3}\times10^{-6}$ M$_\odot$ yr$^{-1}$, which demonstrates that WR 112 is one of the most prolific dust-making WC systems known. △ Less

Submitted 3 August, 2020; originally announced August 2020.

Comments: 17 pages, 9 figures, 1 animated gif, accepted for publication in ApJ

HDR-GAN: HDR Image Reconstruction from Multi-Exposed LDR Images with Large Motions

Authors: Yuzhen Niu, Jianbin Wu, Wenxi Liu, Wenzhong Guo, Rynson W. H. Lau

Abstract: Synthesizing high dynamic range (HDR) images from multiple low-dynamic range (LDR) exposures in dynamic scenes is challenging. There are two major problems caused by the large motions of foreground objects. One is the severe misalignment among the LDR images. The other is the missing content due to the over-/under-saturated regions caused by the moving objects, which may not be easily compensated… ▽ More Synthesizing high dynamic range (HDR) images from multiple low-dynamic range (LDR) exposures in dynamic scenes is challenging. There are two major problems caused by the large motions of foreground objects. One is the severe misalignment among the LDR images. The other is the missing content due to the over-/under-saturated regions caused by the moving objects, which may not be easily compensated for by the multiple LDR exposures. Thus, it requires the HDR generation model to be able to properly fuse the LDR images and restore the missing details without introducing artifacts. To address these two problems, we propose in this paper a novel GAN-based model, HDR-GAN, for synthesizing HDR images from multi-exposed LDR images. To our best knowledge, this work is the first GAN-based approach for fusing multi-exposed LDR images for HDR reconstruction. By incorporating adversarial learning, our method is able to produce faithful information in the regions with missing content. In addition, we also propose a novel generator network, with a reference-based residual merging block for aligning large object motions in the feature domain, and a deep HDR supervision scheme for eliminating artifacts of the reconstructed HDR images. Experimental results demonstrate that our model achieves state-of-the-art reconstruction performance over the prior HDR methods on diverse scenes. △ Less

Submitted 3 July, 2020; originally announced July 2020.

Revisiting the Impact of Dust Production from Carbon-Rich Wolf-Rayet Binaries

Authors: Ryan M. Lau, J. J. Eldridge, Matthew J. Hankins, Astrid Lamberts, Itsuki Sakon, Peredur M. Williams

Abstract: We present a dust spectral energy distribution (SED) and binary stellar population analysis revisiting the dust production rates (DPRs) in the winds of carbon-rich Wolf-Rayet (WC) binaries and their impact on galactic dust budgets. DustEM SED models of 19 Galactic WC ``dustars" reveal DPRs of $\dot{M}_d\sim10^{-10}-10^{-6}$ M$_\odot$ yr$^{-1}$ and carbon dust condensation fractions, $χ_C$, between… ▽ More We present a dust spectral energy distribution (SED) and binary stellar population analysis revisiting the dust production rates (DPRs) in the winds of carbon-rich Wolf-Rayet (WC) binaries and their impact on galactic dust budgets. DustEM SED models of 19 Galactic WC ``dustars" reveal DPRs of $\dot{M}_d\sim10^{-10}-10^{-6}$ M$_\odot$ yr$^{-1}$ and carbon dust condensation fractions, $χ_C$, between $0.002 - 40\%$. A large ($0.1 - 1.0$ $μ$m) dust grain size composition is favored for efficient dustars where $χ_C\gtrsim1\%$. Results for dustars with known orbital periods verify a power-law relation between $χ_C$, orbital period, WC mass-loss rate, and wind velocity consistent with predictions from theoretical models of dust formation in colliding-wind binaries. We incorporated dust production into Binary Population and Spectral Synthesis (BPASS) models to analyze dust production rates from WC dustars, asymptotic giant branch stars (AGBs), red supergiants (RSGs), and core-collapse supernovae (SNe). BPASS models assuming constant star formation (SF) and a co-eval $10^6$ M$_\odot$ stellar population were performed at low, Large Magellanic Cloud (LMC)-like, and solar metallicities (Z = 0.001, 0.008, and 0.020). Both constant SF and co-eval models show that SNe are net dust destroyers at all metallicities. Constant SF models at LMC-like metallicities show that AGB stars slightly outproduce WC binaries and RSGs by factors of $2-3$, whereas at solar metallicites WC binaries are the dominant source of dust for $\sim60$ Myr until the onset of AGBs, which match the dust input of WC binaries. Co-eval population models show that for "bursty" SF, AGB stars dominate dust production at late times ($t\gtrsim 70$ Myr). △ Less

Submitted 15 June, 2020; originally announced June 2020.

Comments: 34 pages, 12 figures, 7 tables, Accepted for publication in ApJ

What makes instance discrimination good for transfer learning?

Authors: Nanxuan Zhao, Zhirong Wu, Rynson W. H. Lau, Stephen Lin

Abstract: Contrastive visual pretraining based on the instance discrimination pretext task has made significant progress. Notably, recent work on unsupervised pretraining has shown to surpass the supervised counterpart for finetuning downstream applications such as object detection and segmentation. It comes as a surprise that image annotations would be better left unused for transfer learning. In this work… ▽ More Contrastive visual pretraining based on the instance discrimination pretext task has made significant progress. Notably, recent work on unsupervised pretraining has shown to surpass the supervised counterpart for finetuning downstream applications such as object detection and segmentation. It comes as a surprise that image annotations would be better left unused for transfer learning. In this work, we investigate the following problems: What makes instance discrimination pretraining good for transfer learning? What knowledge is actually learned and transferred from these models? From this understanding of instance discrimination, how can we better exploit human annotation labels for pretraining? Our findings are threefold. First, what truly matters for the transfer is low-level and mid-level representations, not high-level representations. Second, the intra-category invariance enforced by the traditional supervised model weakens transferability by increasing task misalignment. Finally, supervised pretraining can be strengthened by following an exemplar-based approach without explicit constraints among the instances within the same category. △ Less

Submitted 19 January, 2021; v1 submitted 11 June, 2020; originally announced June 2020.

Comments: Accepted by ICLR 2021

Distilling Localization for Self-Supervised Representation Learning

Authors: Nanxuan Zhao, Zhirong Wu, Rynson W. H. Lau, Stephen Lin

Abstract: Recent progress in contrastive learning has revolutionized unsupervised representation learning. Concretely, multiple views (augmentations) from the same image are encouraged to map to the similar embeddings, while views from different images are pulled apart. In this paper, through visualizing and diagnosing classification errors, we observe that current contrastive models are ineffective at loca… ▽ More Recent progress in contrastive learning has revolutionized unsupervised representation learning. Concretely, multiple views (augmentations) from the same image are encouraged to map to the similar embeddings, while views from different images are pulled apart. In this paper, through visualizing and diagnosing classification errors, we observe that current contrastive models are ineffective at localizing the foreground object, limiting their ability to extract discriminative high-level features. This is due to the fact that view generation process considers pixels in an image uniformly. To address this problem, we propose a data-driven approach for learning invariance to backgrounds. It first estimates foreground saliency in images and then creates augmentations by copy-and-pasting the foreground onto a variety of backgrounds. The learning still follows the instance discrimination pretext task, so that the representation is trained to disregard background content and focus on the foreground. We study a variety of saliency estimation methods, and find that most methods lead to improvements for contrastive learning. With this approach (DiLo), significant performance is achieved for self-supervised learning on ImageNet classification, and also for object detection on PASCAL VOC and MSCOCO. △ Less

Submitted 19 January, 2021; v1 submitted 14 April, 2020; originally announced April 2020.

Comments: Accepted by AAAI2021

Laplacian Denoising Autoencoder

Authors: Jianbo Jiao, Linchao Bao, Yunchao Wei, Shengfeng He, Honghui Shi, Rynson Lau, Thomas S. Huang

Abstract: While deep neural networks have been shown to perform remarkably well in many machine learning tasks, labeling a large amount of ground truth data for supervised training is usually very costly to scale. Therefore, learning robust representations with unlabeled data is critical in relieving human effort and vital for many downstream tasks. Recent advances in unsupervised and self-supervised learni… ▽ More While deep neural networks have been shown to perform remarkably well in many machine learning tasks, labeling a large amount of ground truth data for supervised training is usually very costly to scale. Therefore, learning robust representations with unlabeled data is critical in relieving human effort and vital for many downstream tasks. Recent advances in unsupervised and self-supervised learning approaches for visual data have benefited greatly from domain knowledge. Here we are interested in a more generic unsupervised learning framework that can be easily generalized to other domains. In this paper, we propose to learn data representations with a novel type of denoising autoencoder, where the noisy input data is generated by corrupting latent clean data in the gradient domain. This can be naturally generalized to span multiple scales with a Laplacian pyramid representation of the input data. In this way, the agent learns more robust representations that exploit the underlying data structures across multiple scales. Experiments on several visual benchmarks demonstrate that better representations can be learned with the proposed approach, compared to its counterpart with single-scale corruption and other approaches. Furthermore, we also demonstrate that the learned representations perform well when transferring to other downstream vision tasks. △ Less

Submitted 30 March, 2020; originally announced March 2020.

Night-time Scene Parsing with a Large Real Dataset

Authors: Xin Tan, Ke Xu, Ying Cao, Yiheng Zhang, Lizhuang Ma, Rynson W. H. Lau

Abstract: Although huge progress has been made on scene analysis in recent years, most existing works assume the input images to be in day-time with good lighting conditions. In this work, we aim to address the night-time scene parsing (NTSP) problem, which has two main challenges: 1) labeled night-time data are scarce, and 2) over- and under-exposures may co-occur in the input night-time images and are not… ▽ More Although huge progress has been made on scene analysis in recent years, most existing works assume the input images to be in day-time with good lighting conditions. In this work, we aim to address the night-time scene parsing (NTSP) problem, which has two main challenges: 1) labeled night-time data are scarce, and 2) over- and under-exposures may co-occur in the input night-time images and are not explicitly modeled in existing pipelines. To tackle the scarcity of night-time data, we collect a novel labeled dataset, named {\it NightCity}, of 4,297 real night-time images with ground truth pixel-level semantic annotations. To our knowledge, NightCity is the largest dataset for NTSP. In addition, we also propose an exposure-aware framework to address the NTSP problem through augmenting the segmentation process with explicitly learned exposure features. Extensive experiments show that training on NightCity can significantly improve NTSP performances and that our exposure-aware model outperforms the state-of-the-art methods, yielding top performances on our dataset as well as existing datasets. △ Less

Submitted 1 April, 2022; v1 submitted 15 March, 2020; originally announced March 2020.

Comments: 13 pages, 11 figures. This paper is accepted by IEEE Transactions on Image Processing. The dataset can be accessed via https://dmcv.sjtu.edu.cn/people/phd/tanxin/NightCity/index.html

From SN2010da to NGC 300 ULX-1: Ten Years of Observations of an Unusual High Mass X-ray Binary in NGC 300

Authors: Breanna A. Binder, Stefania Carpano, Marianne Heida, Ryan Lau

Abstract: In 2010 May, an intermediate luminosity optical transient was discovered in the nearby galaxy NGC 300 by a South African amateur astronomer. In the decade since its discovery, multi-wavelength observations of the misnamed ``SN 2010da'' have continually re-shaped our understanding of this high mass X-ray binary system. In this review, we present an overview of the multi-wavelength observations and… ▽ More In 2010 May, an intermediate luminosity optical transient was discovered in the nearby galaxy NGC 300 by a South African amateur astronomer. In the decade since its discovery, multi-wavelength observations of the misnamed ``SN 2010da'' have continually re-shaped our understanding of this high mass X-ray binary system. In this review, we present an overview of the multi-wavelength observations and attempts to understand the 2010 transient event and, later, the re-classification of this system as NGC~300 ULX-1: a red supergiant + neutron star ultraluminous X-ray source. △ Less

Submitted 10 February, 2020; originally announced February 2020.

Comments: Accepted to special edition of Galaxies: High-Mass X-ray Binaries and Ultraluminous X-ray Sources Hosting Neutron Stars

SOFIA/FORCAST Galactic Center Legacy Survey: Overview

Authors: Matthew J. Hankins, Ryan M. Lau, James T. Radomski, Angela S. Cotera, Mark R. Morris, Elisabeth A. C. Mills, Daniel L. Walker, Ashley T. Barnes, Janet P. Simpson, Terry L. Herter, Steven N. Longmore, John Bally, Mansi M. Kasliwal, Nadeen B. Sabha, Macarena Garcia-Marin

Abstract: The Galactic Center contains some of the most extreme conditions for star formation in our Galaxy as well as many other phenomena that are unique to this region. Given our relative proximity to the Galactic Center, we are able to study details of physical processes to a level that is simply not yet possible for more distant galaxies, yielding an otherwise inaccessible view of the nuclear region of… ▽ More The Galactic Center contains some of the most extreme conditions for star formation in our Galaxy as well as many other phenomena that are unique to this region. Given our relative proximity to the Galactic Center, we are able to study details of physical processes to a level that is simply not yet possible for more distant galaxies, yielding an otherwise inaccessible view of the nuclear region of a galaxy. We recently carried out a targeted imaging survey of mid-infrared bright portions of the Galactic Center at 25 and 37 $μ$m using the FORCAST instrument on SOFIA. This survey was one of the inaugural Legacy Programs from SOFIA cycle 7, observing a total area of 403 arcmin$^2$ (2180 pc$^2$), including the Sgr A, B, and C complexes. Here we present an overview of the survey strategy, observations, and data reduction as an accompaniment to the initial public release of the survey data. We discuss interesting regions and features within the data including extended features near the circumnuclear disk, structures in the Arched Filaments and Sickle H II regions, and signs of embedded star formation in Sgr B2 and Sgr C. We also feature a handful of less well studied mid-infrared sources located between Sgr A and Sgr C that could be sites of relatively isolated star formation activity. Last, we discuss plans for subsequent publications and future data releases from the survey. △ Less

Submitted 14 April, 2020; v1 submitted 15 January, 2020; originally announced January 2020.

Comments: 21 pages, 11 figures, Accepted to ApJ

Palomar Gattini-IR: Survey overview, data processing system, on-sky performance and first results

Authors: Kishalay De, Matthew J. Hankins, Mansi M. Kasliwal, Anna M. Moore, Eran O. Ofek, Scott M. Adams, Michael C. B. Ashley, Aliya-Nur Babul, Ashot Bagdasaryan, Kevin B. Burdge, Jill Burnham, Richard G. Dekany, Alexander Declacroix, Antony Galla, Tim Greffe, David Hale, Jacob E. Jencson, Ryan M. Lau, Ashish Mahabal, Daniel McKenna, Manasi Sharma, Patrick L. Shopbell, Roger M. Smith, Jamie Soon, Jennifer Sokoloski , et al. (2 additional authors not shown)

Abstract: (Abridged) Palomar Gattini-IR is a new wide-field, near-infrared robotic time domain survey operating at Palomar Observatory. Using a 30 cm telescope mounted with a H2RG detector, Gattini-IR achieves a field of view of 25 sq. deg. with a pixel scale of 8.7" in J-band. Here, we describe the system design, survey operations, data processing system and on-sky performance of Palomar Gattini-IR. As a p… ▽ More (Abridged) Palomar Gattini-IR is a new wide-field, near-infrared robotic time domain survey operating at Palomar Observatory. Using a 30 cm telescope mounted with a H2RG detector, Gattini-IR achieves a field of view of 25 sq. deg. with a pixel scale of 8.7" in J-band. Here, we describe the system design, survey operations, data processing system and on-sky performance of Palomar Gattini-IR. As a part of the nominal survey, Gattini-IR scans $\approx 7500$ square degrees of the sky every night to a median 5$σ$ depth of $15.7$ AB mag outside the Galactic plane. The survey covers $\approx 15000$ square degrees of the sky visible from Palomar with a median cadence of 2 days. A real-time data processing system produces stacked science images from dithered raw images taken on sky, together with PSF-fit source catalogs and transient candidates identified from subtractions within a median delay of $\approx 4$ hours from the time of observation. The calibrated data products achieve an astrometric accuracy (RMS) of $\approx 0.7$" with respect to Gaia DR2 for sources with S/N $> 10$, and better than $\approx 0.35$" for sources brighter than $\approx 12$ Vega mag. The photometric accuracy (RMS) achieved in the PSF-fit source catalogs is better than $\approx 3$% for sources brighter than $\approx 12$ Vega mag, as calibrated against the 2MASS catalog. With a field of view $\approx 40\times$ larger than any other existing near infrared imaging instrument, Gattini-IR is probing the reddest and dustiest transients in the local universe such as dust obscured supernovae in nearby galaxies, novae behind large columns of extinction within the galaxy, reddened micro-lensing events in the Galactic plane and variability from cool and dust obscured stars. We present results from transients and variables identified since the start of the commissioning period. △ Less

Submitted 29 October, 2019; originally announced October 2019.

Comments: 35 pages, 29 figures. Submitted to PASP

Supernova 2014C: ongoing interaction with extended circumstellar material with silicate dust

Authors: Samaporn Tinyanont, Ryan M Lau, Mansi M Kasliwal, Keiichi Maeda, Nathan Smith, Ori D Fox, Robert D Gehrz, Kishalay De, Jacob Jencson, John Bally, Frank Masci

Abstract: Supernova (SN) 2014C is a unique explosion where a seemingly typical hydrogen-poor stripped envelope SN started to interact with a dense, hydrogen-rich circumstellar medium (CSM) a few months after the explosion. The delayed interaction suggests a detached CSM shell, unlike in a typical SN IIn where the CSM is much closer and the interaction commences earlier post-explosion; indicating a different… ▽ More Supernova (SN) 2014C is a unique explosion where a seemingly typical hydrogen-poor stripped envelope SN started to interact with a dense, hydrogen-rich circumstellar medium (CSM) a few months after the explosion. The delayed interaction suggests a detached CSM shell, unlike in a typical SN IIn where the CSM is much closer and the interaction commences earlier post-explosion; indicating a different mass loss history. We present near- to mid-infrared observations of SN 2014C from 1-5 years after the explosion, including uncommon 9.7 $μ$m imaging with COMICS on the Subaru telescope. Spectroscopy shows that the interaction is still ongoing, with the intermediate-width He I 1.083 $μ$m emission present out to our latest epoch 1639 days post-explosion. The last Spitzer/IRAC photometry at 1920 days post-explosion further confirms ongoing CSM interaction. The 1-10 $μ$m spectral energy distributions (SEDs) can be explained by a dust model with a mixture of 69% carbonaceous and 31% silicate dust, pointing to a chemically inhomogeneous CSM. The inference of silicate dust is the first among interacting SNe. An SED model with purely carbonaceous CSM dust is possible, but would require more than 0.22 $M_{\odot}$ of dust, which is an order of magnitude larger than what observed in any other SNe, measured in the same way, at this epoch. The light curve beyond 500 days is well fit by an interaction model with a wind-driven CSM and a mass loss rate of $\sim 10^{-3} \, M_{\odot}\,\rm yr^{-1}$, which presents an additional CSM density component exterior to the constant density shell reported previously in the literature. SN 2014C could originate in a binary system, similar to RY Scuti, which would explain the observed chemical and density profile inhomogeneity in the CSM. △ Less

Submitted 13 September, 2019; originally announced September 2019.

Comments: Submitted to ApJ

Discovery of a red supergiant donor star in SN2010da/NGC 300 ULX-1

Authors: M. Heida, R. M. Lau, B. Davies, M. Brightman, F. Fürst, B. W. Grefenstette, J. A. Kennea, F. Tramper, D. J. Walton, F. A. Harrison

Abstract: SN2010da/NGC 300 ULX-1 was first detected as a supernova impostor in May 2010 and was recently discovered to be a pulsating ultraluminous X-ray source. In this letter, we present VLT/X-shooter spectra of this source obtained in October 2018, covering the wavelength range 350-2300 nm. The $J$- and $H$-bands clearly show the presence of a red supergiant donor star that is best matched by a MARCS ste… ▽ More SN2010da/NGC 300 ULX-1 was first detected as a supernova impostor in May 2010 and was recently discovered to be a pulsating ultraluminous X-ray source. In this letter, we present VLT/X-shooter spectra of this source obtained in October 2018, covering the wavelength range 350-2300 nm. The $J$- and $H$-bands clearly show the presence of a red supergiant donor star that is best matched by a MARCS stellar atmosphere with $T_{\rm eff} = 3650 - 3900$ K and $\log(L_{\rm bol}/L_{\odot}) = 4.25\pm0.10$, which yields a stellar radius $R = 310 \pm 70 R_{\odot}$. To fit the full spectrum, two additional components are required: a blue excess that can be fitted either by a hot blackbody (T $\gtrsim 20,000$ K) or a power law (spectral index $α\approx 4$) and is likely due to X-ray emission reprocessed in the outer accretion disk or the donor star; and a red excess that is well fitted by a blackbody with a temperature of $\sim 1100$ K, and is likely due to warm dust in the vicinity of SN2010da. The presence of a red supergiant in this system implies an orbital period of at least 0.8-2.1 years, assuming Roche lobe overflow. Given the large donor-to-compact object mass ratio, orbital modulations of the radial velocity of the red supergiant are likely undetectable. However, the radial velocity amplitude of the neutron star is large enough (up to 40-60 km s$^{-1}$) to potentially be measured in the future, unless the system is viewed at a very unfavorable inclination. △ Less

Submitted 4 September, 2019; originally announced September 2019.

Comments: 10 pages, 6 figures, accepted for publication in ApJ Letters

Dual Student: Breaking the Limits of the Teacher in Semi-supervised Learning

Authors: Zhanghan Ke, Daoye Wang, Qiong Yan, Jimmy Ren, Rynson W. H. Lau

Abstract: Recently, consistency-based methods have achieved state-of-the-art results in semi-supervised learning (SSL). These methods always involve two roles, an explicit or implicit teacher model and a student model, and penalize predictions under different perturbations by a consistency constraint. However, the weights of these two roles are tightly coupled since the teacher is essentially an exponential… ▽ More Recently, consistency-based methods have achieved state-of-the-art results in semi-supervised learning (SSL). These methods always involve two roles, an explicit or implicit teacher model and a student model, and penalize predictions under different perturbations by a consistency constraint. However, the weights of these two roles are tightly coupled since the teacher is essentially an exponential moving average (EMA) of the student. In this work, we show that the coupled EMA teacher causes a performance bottleneck. To address this problem, we introduce Dual Student, which replaces the teacher with another student. We also define a novel concept, stable sample, following which a stabilization constraint is designed for our structure to be trainable. Further, we discuss two variants of our method, which produce even higher performance. Extensive experiments show that our method improves the classification performance significantly on several main SSL benchmarks. Specifically, it reduces the error rate of the 13-layer CNN from 16.84% to 12.39% on CIFAR-10 with 1k labels and from 34.10% to 31.56% on CIFAR-100 with 10k labels. In addition, our method also achieves a clear improvement in domain adaptation. △ Less

Submitted 3 September, 2019; originally announced September 2019.

Comments: International Conference in Computer Vision 2019 (ICCV 2019)

Where Is My Mirror?

Authors: Xin Yang, Haiyang Mei, Ke Xu, Xiaopeng Wei, Baocai Yin, Rynson W. H. Lau

Abstract: Mirrors are everywhere in our daily lives. Existing computer vision systems do not consider mirrors, and hence may get confused by the reflected content inside a mirror, resulting in a severe performance degradation. However, separating the real content outside a mirror from the reflected content inside it is non-trivial. The key challenge is that mirrors typically reflect contents similar to thei… ▽ More Mirrors are everywhere in our daily lives. Existing computer vision systems do not consider mirrors, and hence may get confused by the reflected content inside a mirror, resulting in a severe performance degradation. However, separating the real content outside a mirror from the reflected content inside it is non-trivial. The key challenge is that mirrors typically reflect contents similar to their surroundings, making it very difficult to differentiate the two. In this paper, we present a novel method to segment mirrors from an input image. To the best of our knowledge, this is the first work to address the mirror segmentation problem with a computational approach. We make the following contributions. First, we construct a large-scale mirror dataset that contains mirror images with corresponding manually annotated masks. This dataset covers a variety of daily life scenes, and will be made publicly available for future research. Second, we propose a novel network, called MirrorNet, for mirror segmentation, by modeling both semantical and low-level color/texture discontinuities between the contents inside and outside of the mirrors. Third, we conduct extensive experiments to evaluate the proposed method, and show that it outperforms the carefully chosen baselines from the state-of-the-art detection and segmentation methods. △ Less

Submitted 3 October, 2019; v1 submitted 24 August, 2019; originally announced August 2019.

Comments: Accepted by ICCV 2019. Project homepage: https://mhaiyang.github.io/ICCV2019_MirrorNet/index.html

Uncovering Red and Dusty Ultraluminous X-ray Sources with Spitzer

Authors: Ryan M. Lau, Marianne Heida, Dominic J. Walton, Mansi M. Kasliwal, Scott M. Adams, Ann Marie Cody, Kishalay De, Robert D. Gehrz, Felix Furst, Jacob E. Jencson, jamie A. Kennea, Frank Masci

Abstract: We present a mid-infrared (IR) sample study of nearby ultraluminous X-ray sources (ULXs) using multi-epoch observations with the Infrared Array Camera (IRAC) on the Spitzer Space Telescope. Spitzer/IRAC observations taken after 2014 were obtained as part of the Spitzer Infrared Intensive Transients Survey (SPIRITS). Our sample includes 96 ULXs located within 10 Mpc. Of the 96~ULXs, 12 have candida… ▽ More We present a mid-infrared (IR) sample study of nearby ultraluminous X-ray sources (ULXs) using multi-epoch observations with the Infrared Array Camera (IRAC) on the Spitzer Space Telescope. Spitzer/IRAC observations taken after 2014 were obtained as part of the Spitzer Infrared Intensive Transients Survey (SPIRITS). Our sample includes 96 ULXs located within 10 Mpc. Of the 96~ULXs, 12 have candidate counterparts consistent with absolute mid-IR magnitudes of supergiants, and 16 counterparts exceeded the mid-IR brightness of single supergiants and are thus more consistent with star clusters or non-ULX background active galactic nuclei (AGN). The supergiant candidate counterparts exhibit a bi-modal color distribution in a Spitzer/IRAC color-magnitude diagram, where "red" and "blue" ULXs fall in IRAC colors $[3.6] - [4.5]\sim0.7$ and $[3.6] - [4.5]\sim0.0$, respectively. The mid-IR colors and absolute magnitudes of 4 "red" and 5 "blue" ULXs are consistent with that of supergiant B[e] (sgB[e]) and red supergiant (RSG) stars, respectively. While "blue", RSG-like mid-IR ULX counterparts likely host RSG mass donors, we propose the "red" counterparts are ULXs exhibiting the "B[e] phenomenon'' rather than hosts of sgB[e] mass donors. We show that the mid-IR excess from the "red" ULXs is likely due to thermal emission from circumstellar or circumbinary dust. Using dust as a probe for total mass, we estimate mass-loss rates of $\dot{M}\sim1\times10^{-4}$ M$_\odot$ yr$^{-1}$ in dust-forming outflows of red ULXs. Based on the transient mid-IR behavior and its relatively flat spectral index, $α=-0.19\pm0.1$, we suggest that the mid-IR emission from Holmberg IX X-1 originates from a variable jet. △ Less

Submitted 18 April, 2019; originally announced April 2019.

Comments: 40 pages, 11 figures, accepted for publication in ApJ

Discovery of an intermediate-luminosity red transient in M51 and its likely dust-obscured, infrared-variable progenitor

Authors: Jacob E. Jencson, Scott M. Adams, Howard E. Bond, Schuyler D. van Dyk, Mansi M. Kasliwal, John Bally, Nadejda Blagorodnova, Kishalay De, Christoffer Fremling, Yuhan Yao, Andrew Fruchter, David Rubin, Cristina Barbarino, Jesper Sollerman, Adam A. Miller, Erin K. S. Hicks, Matthew A. Malkan, Igor Andreoni, Eric C. Bellm, Robert Buchheim, Richard Dekany, Michael Feeney, Sara Frederick, Avishay Gal-Yam, Robert D. Gehrz , et al. (27 additional authors not shown)

Abstract: We present the discovery of an optical transient (OT) in Messier 51, designated M51 OT2019-1 (also ZTF19aadyppr, AT 2019abn, ATLAS19bzl), by the Zwicky Transient Facility (ZTF). The OT rose over 15 days to an observed luminosity of $M_r=-13$ ($νL_ν=9\times10^6~L_{\odot}$), in the luminosity gap between novae and typical supernovae (SNe). Spectra during the outburst show a red continuum, Balmer emi… ▽ More We present the discovery of an optical transient (OT) in Messier 51, designated M51 OT2019-1 (also ZTF19aadyppr, AT 2019abn, ATLAS19bzl), by the Zwicky Transient Facility (ZTF). The OT rose over 15 days to an observed luminosity of $M_r=-13$ ($νL_ν=9\times10^6~L_{\odot}$), in the luminosity gap between novae and typical supernovae (SNe). Spectra during the outburst show a red continuum, Balmer emission with a velocity width of $\approx400$ km s$^{-1}$, Ca II and [Ca II] emission, and absorption features characteristic of an F-type supergiant. The spectra and multiband light curves are similar to the so-called "SN impostors" and intermediate-luminosity red transients (ILRTs). We directly identify the likely progenitor in archival Spitzer Space Telescope imaging with a $4.5~μ$m luminosity of $M_{[4.5]}\approx-12.2$ and a $[3.6]-[4.5]$ color redder than 0.74 mag, similar to those of the prototype ILRTs SN 2008S and NGC 300 OT2008-1. Intensive monitoring of M51 with Spitzer further reveals evidence for variability of the progenitor candidate at [4.5] in the years before the OT. The progenitor is not detected in pre-outburst Hubble Space Telescope optical and near-IR images. The optical colors during outburst combined with spectroscopic temperature constraints imply a higher reddening of $E(B-V)\approx0.7$ mag and higher intrinsic luminosity of $M_r\approx-14.9$ ($νL_ν=5.3\times10^7~L_{\odot}$) near peak than seen in previous ILRT candidates. Moreover, the extinction estimate is higher on the rise than on the plateau, suggestive of an extended phase of circumstellar dust destruction. These results, enabled by the early discovery of M51 OT2019-1 and extensive pre-outburst archival coverage, offer new clues about the debated origins of ILRTs and may challenge the hypothesis that they arise from the electron-capture induced collapse of extreme asymptotic giant branch stars. △ Less

Submitted 29 July, 2019; v1 submitted 15 April, 2019; originally announced April 2019.

Comments: 21 pages, 5 figures, published in ApJL

Journal ref: The Astrophysical Journal Letters, 880 (2019) L20

Spatial Attentive Single-Image Deraining with a High Quality Real Rain Dataset

Authors: Tianyu Wang, Xin Yang, Ke Xu, Shaozhe Chen, Qiang Zhang, Rynson Lau

Abstract: Removing rain streaks from a single image has been drawing considerable attention as rain streaks can severely degrade the image quality and affect the performance of existing outdoor vision tasks. While recent CNN-based derainers have reported promising performances, deraining remains an open problem for two reasons. First, existing synthesized rain datasets have only limited realism, in terms of… ▽ More Removing rain streaks from a single image has been drawing considerable attention as rain streaks can severely degrade the image quality and affect the performance of existing outdoor vision tasks. While recent CNN-based derainers have reported promising performances, deraining remains an open problem for two reasons. First, existing synthesized rain datasets have only limited realism, in terms of modeling real rain characteristics such as rain shape, direction and intensity. Second, there are no public benchmarks for quantitative comparisons on real rain images, which makes the current evaluation less objective. The core challenge is that real world rain/clean image pairs cannot be captured at the same time. In this paper, we address the single image rain removal problem in two ways. First, we propose a semi-automatic method that incorporates temporal priors and human supervision to generate a high-quality clean image from each input sequence of real rain images. Using this method, we construct a large-scale dataset of $\sim$$29.5K$ rain/rain-free image pairs that covers a wide range of natural rain scenes. Second, to better cover the stochastic distribution of real rain streaks, we propose a novel SPatial Attentive Network (SPANet) to remove rain streaks in a local-to-global manner. Extensive experiments demonstrate that our network performs favorably against the state-of-the-art deraining methods. △ Less

Submitted 18 August, 2019; v1 submitted 2 April, 2019; originally announced April 2019.

Comments: Accepted by CVPR'19. Project page: https://stevewongv.github.io/derain-project.html

The Dynamic Infrared Sky

Authors: Mansi M. Kasliwal, Scott Adams, Igor Andreoni, Michael Ashley, Nadia Blagorodnova, Kishalay De, Danielle Frostig, Gabor Furesz, Jacob Jencson, Matthew Hankins, George Helou, Ryan Lau, Anna Moore, Eran Ofek, Rob Simcoe, Jennifer Sokoloski, Jamie Soon, Samaporn Tinyanont, Tony Travouillon

Abstract: Opening up the dynamic infrared sky for systematic time-domain exploration would yield many scientific advances. Multi-messenger pursuits such as localizing gravitational waves from neutron star mergers and quantifying the nucleosynthetic yields require the infrared. Another multi-messenger endeavor that needs infrared surveyors is the study of the much-awaited supernova in our own Milky Way. Unde… ▽ More Opening up the dynamic infrared sky for systematic time-domain exploration would yield many scientific advances. Multi-messenger pursuits such as localizing gravitational waves from neutron star mergers and quantifying the nucleosynthetic yields require the infrared. Another multi-messenger endeavor that needs infrared surveyors is the study of the much-awaited supernova in our own Milky Way. Understanding shocks in novae, true rates of supernovae and stellar mergers are some other examples of stellar evolution and high energy physics wherein the answers are buried in the infrared. We discuss some of the challenges in the infrared and pathfinders to overcome them. We conclude with recommendations on both infrared discovery engines and infrared follow-up machines that would enable this field to flourish in the next decade. △ Less

Submitted 19 March, 2019; originally announced March 2019.

Comments: Astro2020 Science White Paper for Decadal Survey

SPIRITS Catalog of Infrared Variables: Identification of Extremely Luminous Long Period Variables

Authors: V. R. Karambelkar, S. M. Adams, P. A. Whitelock, M. M. Kasliwal, J. E. Jencson, M. L. Boyer, S. R. Goldman, F. Masci, A. M. Cody, J. Bally, H. E. Bond, R. D. Gehrz, M. Parthasarathy, R. M. Lau

Abstract: We present a catalog of 417 luminous infrared variable stars with periods exceeding 250 days. These were identified in 20 nearby galaxies by the ongoing SPIRITS survey with the Spitzer Space Telescope. Of these, 359 variables have $M_{[4.5]}$ (phase-weighted mean magnitudes) fainter than $-12$ and periods and luminosities consistent with previously reported variables in the Large Magellanic Cloud.… ▽ More We present a catalog of 417 luminous infrared variable stars with periods exceeding 250 days. These were identified in 20 nearby galaxies by the ongoing SPIRITS survey with the Spitzer Space Telescope. Of these, 359 variables have $M_{[4.5]}$ (phase-weighted mean magnitudes) fainter than $-12$ and periods and luminosities consistent with previously reported variables in the Large Magellanic Cloud. However, 58 variables are more luminous than $M_{[4.5]} = -12$, including 11 that are brighter than $M_{[4.5]} = -13$ with the brightest having $M_{[4.5]} = -15.51$. Most of these bright variable sources have quasi-periods longer than 1000 days, including four over 2000 days. We suggest that the fundamental period-luminosity relationship, previously measured for the Large Magellanic Cloud, extends to much higher luminosities and longer periods in this large galaxy sample. We posit that these variables include massive AGB stars (possibly super-AGB stars), red supergiants experiencing exceptionally high mass-loss rates, and interacting binaries. We also present 3.6, 4.5, 5.8 and 8.0 $μ$m photometric catalogs for all sources in these 20 galaxies. △ Less

Submitted 6 May, 2019; v1 submitted 22 January, 2019; originally announced January 2019.

Comments: 18 pages, 25 figures

Supernova 2017eaw: molecule and dust formation from infrared observations

Authors: Samaporn Tinyanont, Mansi M Kasliwal, Kelsie Krafton, Ryan Lau, Jeonghee Rho, Douglas C Leonard, Kishalay De, Jacob Jencson, Dimitri Mawet, Maxwell Millar-Blanchaer, Ricky Nilsson, Lin Yan, Robert D Gehrz, George Helou, Schuyler D Van Dyk, Eugene Serabyn, Ori D Fox, Geoffrey Clayton

Abstract: We present infrared (IR) photometry and spectroscopy of the Type II-P SN 2017eaw and its progenitor in the nearby galaxy NGC 6946. Progenitor observations in the Ks band in 4 epochs from 1 year to 1 day before the explosion reveal no significant variability in the progenitor star greater than 6% that last longer than 200 days. SN 2017eaw is a typical SN II-P with near-IR and mid-IR photometric evo… ▽ More We present infrared (IR) photometry and spectroscopy of the Type II-P SN 2017eaw and its progenitor in the nearby galaxy NGC 6946. Progenitor observations in the Ks band in 4 epochs from 1 year to 1 day before the explosion reveal no significant variability in the progenitor star greater than 6% that last longer than 200 days. SN 2017eaw is a typical SN II-P with near-IR and mid-IR photometric evolution similar to those of SNe 2002hh and 2004et, other normal SNe II-P in the same galaxy. Spectroscopic monitoring between 389 and 480 days post explosion reveals strong CO first overtone emission at 389 d, with a line profile matching that of SN 1987A from the same epoch, indicating $\sim 10^{-3} \, M_{\odot}$ of CO at 1,800 K. From the 389 d epoch until the most recent observation at 566 d, the first overtone feature fades while the 4.5 $μ$m excess, likely from the CO fundamental band, remains. This behavior indicates that the CO has not been destroyed, but that the gas has cooled enough that the levels responsible for first overtone emissions are no longer populated. Finally, the evolution of Spitzer 3.6 $μ$m photometry shows evidence for dust formation in SN 2017eaw, with a dust mass of $10^{-6}$ or $10^{-4}\,M_{\odot}$ assuming carbonaceous or silicate grains respectively. △ Less

Submitted 7 January, 2019; originally announced January 2019.

The SPIRITS sample of Luminous Infrared Transients: Uncovering Hidden Supernovae and Dusty Stellar Outbursts in Nearby Galaxies

Authors: Jacob E. Jencson, Mansi M. Kasliwal, Scott M. Adams, Howard E. Bond, Kishalay De, Joel Johansson, Viraj Karambelkar, Ryan M. Lau, Samaporn Tinyanont, Stuart D. Ryder, Ann Marie Cody, Frank J. Masci, John Bally, Nadia Blagorodnova, Sergio Castellón, Christoffer Fremling, Robert D. Gehrz, George Helou, Charles D. Kilpatrick, Peter A. Milne, Nidia Morrell, Daniel A. Perley, M. M. Phillips, Nathan Smith, Schuyler D. van Dyk , et al. (1 additional authors not shown)

Abstract: We present a systematic study of the most luminous ($M_{\mathrm{IR}}$ [Vega magnitudes] brighter than $-14$) infrared (IR) transients discovered by the SPitzer InfraRed Intensive Transients Survey (SPIRITS) between 2014 and 2018 in nearby galaxies ($D < 35$ Mpc). The sample consists of nine events that span peak IR luminosities of $M_{[4.5],\mathrm{peak}}$ between $-14$ and $-18.2$, show IR colors… ▽ More We present a systematic study of the most luminous ($M_{\mathrm{IR}}$ [Vega magnitudes] brighter than $-14$) infrared (IR) transients discovered by the SPitzer InfraRed Intensive Transients Survey (SPIRITS) between 2014 and 2018 in nearby galaxies ($D < 35$ Mpc). The sample consists of nine events that span peak IR luminosities of $M_{[4.5],\mathrm{peak}}$ between $-14$ and $-18.2$, show IR colors between $0.2 < ([3.6]{-}[4.5]) < 3.0$, and fade on timescales between $55$ days $< t_{\mathrm{fade}} < 480$ days. The two reddest events ($A_V > 12$) show multiple, luminous IR outbursts over several years and have directly detected, massive progenitors in archival imaging. With analyses of extensive, multiwavelength follow-up, we suggest the following possible classifications: five obscured core-collapse supernovae (CCSNe), two erupting massive stars, one luminous red nova, and one intermediate-luminosity red transient. We define a control sample of all optically discovered transients recovered in SPIRITS galaxies and satisfying the same selection criteria. The control sample consists of eight CCSNe and one Type Iax SN. We find that 7 of the 13 CCSNe in the SPIRITS sample have lower bounds on their extinction of $2 < A_V < 8$. We estimate a nominal fraction of CCSNe in nearby galaxies that are missed by optical surveys as high as $38.5^{+26.0}_{-21.9}$% (90% confidence). This study suggests that a significant fraction of CCSNe may be heavily obscured by dust and therefore undercounted in the census of nearby CCSNe from optical searches. △ Less

Submitted 6 January, 2020; v1 submitted 3 January, 2019; originally announced January 2019.

Comments: 47 pages, 16 figures, published in ApJ

Journal ref: The Astrophysical Journal 886 (2019) 40