Showing 1–6 of 6 results for author: Dwyer, K

Contrastive Self-Supervised Learning for Spatio-Temporal Analysis of Lung Ultrasound Videos

Authors: Li Chen, Jonathan Rubin, Jiahong Ouyang, Naveen Balaraju, Shubham Patil, Courosh Mehanian, Sourabh Kulhare, Rachel Millin, Kenton W Gregory, Cynthia R Gregory, Meihua Zhu, David O Kessler, Laurie Malia, Almaz Dessie, Joni Rabiner, Di Coneybeare, Bo Shopsin, Andrew Hersh, Cristian Madar, Jeffrey Shupp, Laura S Johnson, Jacob Avila, Kristin Dwyer, Peter Weimersheimer, Balasundar Raju , et al. (2 additional authors not shown)

Abstract: Self-supervised learning (SSL) methods have shown promise for medical imaging applications by learning meaningful visual representations, even when the amount of labeled data is limited. Here, we extend state-of-the-art contrastive learning SSL methods to 2D+time medical ultrasound video data by introducing a modified encoder and augmentation method capable of learning meaningful spatio-temporal r… ▽ More Self-supervised learning (SSL) methods have shown promise for medical imaging applications by learning meaningful visual representations, even when the amount of labeled data is limited. Here, we extend state-of-the-art contrastive learning SSL methods to 2D+time medical ultrasound video data by introducing a modified encoder and augmentation method capable of learning meaningful spatio-temporal representations, without requiring constraints on the input data. We evaluate our method on the challenging clinical task of identifying lung consolidations (an important pathological feature) in ultrasound videos. Using a multi-center dataset of over 27k lung ultrasound videos acquired from over 500 patients, we show that our method can significantly improve performance on downstream localization and classification of lung consolidation. Comparisons against baseline models trained without SSL show that the proposed methods are particularly advantageous when the size of labeled training data is limited (e.g., as little as 5% of the training set). △ Less

Submitted 14 October, 2023; originally announced October 2023.

Comments: ISBI 2023, 2023 IEEE 20th International Symposium on Biomedical Imaging (ISBI)

Weakly Semi-Supervised Detection in Lung Ultrasound Videos

Authors: Jiahong Ouyang, Li Chen, Gary Y. Li, Naveen Balaraju, Shubham Patil, Courosh Mehanian, Sourabh Kulhare, Rachel Millin, Kenton W. Gregory, Cynthia R. Gregory, Meihua Zhu, David O. Kessler, Laurie Malia, Almaz Dessie, Joni Rabiner, Di Coneybeare, Bo Shopsin, Andrew Hersh, Cristian Madar, Jeffrey Shupp, Laura S. Johnson, Jacob Avila, Kristin Dwyer, Peter Weimersheimer, Balasundar Raju , et al. (2 additional authors not shown)

Abstract: Frame-by-frame annotation of bounding boxes by clinical experts is often required to train fully supervised object detection models on medical video data. We propose a method for improving object detection in medical videos through weak supervision from video-level labels. More concretely, we aggregate individual detection predictions into video-level predictions and extend a teacher-student train… ▽ More Frame-by-frame annotation of bounding boxes by clinical experts is often required to train fully supervised object detection models on medical video data. We propose a method for improving object detection in medical videos through weak supervision from video-level labels. More concretely, we aggregate individual detection predictions into video-level predictions and extend a teacher-student training strategy to provide additional supervision via a video-level loss. We also introduce improvements to the underlying teacher-student framework, including methods to improve the quality of pseudo-labels based on weak supervision and adaptive schemes to optimize knowledge transfer between the student and teacher networks. We apply this approach to the clinically important task of detecting lung consolidations (seen in respiratory infections such as COVID-19 pneumonia) in medical ultrasound videos. Experiments reveal that our framework improves detection accuracy and robustness compared to baseline semi-supervised models, and improves efficiency in data and annotation usage. △ Less

Submitted 7 August, 2023; originally announced August 2023.

Comments: IPMI 2023

Reaction pathways of BCl$_3$ for acceptor delta-do** of silicon

Authors: Quinn Campbell, Kevin J. Dwyer, Sungha Baek, Andrew D. Baczewski, Robert E. Butera, Shashank Misra

Abstract: BCl$_3$ is a promising candidate for atomic-precision acceptor do** in Si, but optimizing the electrical properties of structures created with this technique requires a detailed understanding of adsorption and dissociation pathways for this precursor. Here, we use density functional theory and scanning tunneling microscopy (STM) to identify and explore these pathways for BCl$_3$ on Si(100) at di… ▽ More BCl$_3$ is a promising candidate for atomic-precision acceptor do** in Si, but optimizing the electrical properties of structures created with this technique requires a detailed understanding of adsorption and dissociation pathways for this precursor. Here, we use density functional theory and scanning tunneling microscopy (STM) to identify and explore these pathways for BCl$_3$ on Si(100) at different annealing temperatures. We demonstrate that BCl$_3$ adsorbs selectively without a reaction barrier, and subsequently dissociates relatively easily with reaction barriers $\approx$1 eV. Using this dissociation pathway, we parameterize a Kinetic Monte Carlo model to predict B incorporation rates as a function of dosing conditions. STM is used to image BCl$_{3}$ adsorbates, identifying several surface configurations and tracking the change in their distribution as a function of the annealing temperature, matching predictions of the kinetic model well. This straightforward pathway for atomic-precision acceptor do** helps enable a wide range of applications including bipolar nanoelectronics, acceptor-based qubits, and superconducting Si. △ Less

Submitted 27 January, 2022; originally announced January 2022.

Comments: 20 pages, 5 figures, Main text + supporting info

Area-selective deposition and B $δ$-do** of Si(100) with BCl$_{3}$

Authors: K. J. Dwyer, S. Baek, Azadeh Farzaneh, Michael Dreyer, J. R. Williams, R. E. Butera

Abstract: B-doped $δ$-layers were fabricated in Si(100) using BCl$_{3}$ as a dopant precursor in ultrahigh vacuum. BCl$_{3}$ adsorbed readily at room temperature, as revealed by scanning tunneling microscopy (STM) imaging. Annealing at elevated temperatures facilitated B incorporation into the Si substrate. Secondary ion mass spectrometry (SIMS) depth profiling demonstrated a peak B concentration $>$ 1.2(1)… ▽ More B-doped $δ$-layers were fabricated in Si(100) using BCl$_{3}$ as a dopant precursor in ultrahigh vacuum. BCl$_{3}$ adsorbed readily at room temperature, as revealed by scanning tunneling microscopy (STM) imaging. Annealing at elevated temperatures facilitated B incorporation into the Si substrate. Secondary ion mass spectrometry (SIMS) depth profiling demonstrated a peak B concentration $>$ 1.2(1) $\times$ 10$^{21}$ cm$^{-3}$ with a total areal dose of 1.85(1) $\times$ 10$^{14}$ cm$^{-2}$ resulting from a 30 L BCl$_{3}$ dose at 150 $^{\circ}$C. Hall bar measurements of a similar sample were performed at 3.0 K revealing a sheet resistance of $R_{\mathrm{s}}$ = 1.91 k$Ω\square^{-1}$, a hole concentration of $n$ = 1.90 $\times$ 10$^{14}$ cm$^{-2}$ and a hole mobility of $μ$ = 38.0 cm$^{2}$V$^{-1}$s$^{-1}$ without performing an incorporation anneal. Further, the conductivity of several B-doped $δ$-layers showed a log dependence on temperature suggestive of a two-dimensional system. Selective-area deposition of BCl$_{3}$ was also demonstrated using both H- and Cl-based monatomic resists. In comparison to a dosed area on bare Si, adsorption selectivity ratios for H and Cl resists were determined by SIMS to be 310(10):1 and 1529(5):1, respectively, further validating the use of BCl$_{3}$ as a dopant precursor for atomic precision fabrication of acceptor-doped devices in Si. △ Less

Submitted 12 March, 2021; originally announced March 2021.

AlCl$_{3}$-dosed Si(100)-2$\times$1: Adsorbates, chlorinated Al chains, and incorporated Al

Authors: Matthew S. Radue, Sungha Baek, Azadeh Farzaneh, K. J. Dwyer, Quinn Campbell, Andrew D. Baczewski, Ezra Bussmann, George T. Wang, Yifei Mo, Shashank Misra, R. E. Butera

Abstract: The adsorption of AlCl$_{3}$ on Si(100) and the effect of annealing the AlCl$_{3}$-dosed substrate was studied to reveal key surface processes for the development of atomic-precision acceptor-do** techniques. This investigation was performed via scanning tunneling microscopy (STM), X-ray photoelectron spectroscopy (XPS), and density functional theory (DFT) calculations. At room temperature, AlCl… ▽ More The adsorption of AlCl$_{3}$ on Si(100) and the effect of annealing the AlCl$_{3}$-dosed substrate was studied to reveal key surface processes for the development of atomic-precision acceptor-do** techniques. This investigation was performed via scanning tunneling microscopy (STM), X-ray photoelectron spectroscopy (XPS), and density functional theory (DFT) calculations. At room temperature, AlCl$_{3}$ readily adsorbed to the Si substrate dimers and dissociated to form a variety of species. Annealing of the AlCl$_{3}$-dosed substrate at temperatures below 450 $^{\circ}$C produced unique chlorinated aluminum chains (CACs) elongated along the Si(100) dimer row direction. An atomic model for the chains is proposed with supporting DFT calculations. Al was incorporated into the Si substrate upon annealing at 450 $^{\circ}$C and above, and Cl desorption was observed for temperatures beyond 450 $^{\circ}$C. Al-incorporated samples were encapsulated in Si and characterized by secondary ion mass spectrometry (SIMS) depth profiling to quantify the Al atom concentration, which was found to be in excess of 10$^{20}$ cm$^{-3}$ across a $\sim$2.7 nm thick $δ$-doped region. The Al concentration achieved here and the processing parameters utilized promote AlCl$_{3}$ as a viable gaseous precursor for novel acceptor-doped Si materials and devices for quantum computing. △ Less

Submitted 22 January, 2021; originally announced January 2021.

Comments: 35 pages, 8 figures

STM-induced desorption and lithographic patterning of Cl-Si(100)-(2x1)

Authors: K. J Dwyer, Michael Dreyer, R. E. Butera

Abstract: We investigated STM-induced chlorine desorption and lithographic patterning of Cl-terminated Si(100)-(2x1) surfaces at sample temperatures from 4 K to 600 K. STM lithography has previously focused on hydrogen-based chemistry for donor device fabrication. Here, to develop halogen-based chemistries for fabricating acceptor-based devices, we substituted the hydrogen resist with chlorine. Lithographic… ▽ More We investigated STM-induced chlorine desorption and lithographic patterning of Cl-terminated Si(100)-(2x1) surfaces at sample temperatures from 4 K to 600 K. STM lithography has previously focused on hydrogen-based chemistry for donor device fabrication. Here, to develop halogen-based chemistries for fabricating acceptor-based devices, we substituted the hydrogen resist with chlorine. Lithographic patterning was explored using both field emission patterning to desorb chlorine from large areas as well as atomic precision patterning to desorb chlorine along one to two dimer rows at a time. We varied the experimental parameters for lithographic patterning and found a positive correlation between pattern line widths and both sample bias voltage and total electron dose. Finally, the use of chlorine, bromine, and iodine as lithographic resists to broaden the range of available chemistries for future device fabrication utilizing halogen-based dopant precursors is discussed. △ Less

Submitted 14 July, 2019; v1 submitted 16 August, 2018; originally announced August 2018.

Comments: 10 pages, 8 figures