-
Small-scale cosmic ray anisotropy observed by the GRAPES-3 experiment at TeV energies
Authors:
M. Chakraborty,
S. Ahmad,
A. Chandra,
S. R. Dugad,
U. D. Goswami,
S. K. Gupta,
B. Hariharan,
Y. Hayashi,
P. Jagadeesan,
A. Jain,
P. Jain,
S. Kawakami,
T. Koi,
H. Kojima,
S. Mahapatra,
P. K. Mohanty,
R. Moharana,
Y. Muraki,
P. K. Nayak,
T. Nonaka,
T. Nakamura,
A. Oshima,
B. P. Pant,
D. Pattanaik,
S. Paul
, et al. (13 additional authors not shown)
Abstract:
GRAPES-3 is a mid-altitude (2200 m) and near equatorial ($11.4^{\circ}$ North) air shower array, overlap** in its field of view for cosmic ray observations with experiments that are located in Northern and Southern hemispheres. We analyze a sample of $3.7\times10^9$ cosmic ray events collected by the GRAPES-3 experiment between 1 January 2013 and 31 December 2016 with a median energy of…
▽ More
GRAPES-3 is a mid-altitude (2200 m) and near equatorial ($11.4^{\circ}$ North) air shower array, overlap** in its field of view for cosmic ray observations with experiments that are located in Northern and Southern hemispheres. We analyze a sample of $3.7\times10^9$ cosmic ray events collected by the GRAPES-3 experiment between 1 January 2013 and 31 December 2016 with a median energy of $\sim16$ TeV for study of small-scale ($<60^{\circ}$) angular scale anisotropies. We observed two structures labeled as A and B, deviate from the expected isotropic distribution of cosmic rays in a statistically significant manner. Structure `A' spans $50^{\circ}$ to $80^{\circ}$ in the right ascension and $-15^{\circ}$ to $30^{\circ}$ in the declination coordinate. The relative excess observed in the structure A is at the level of $(6.5\pm1.3)\times10^{-4}$ with a statistical significance of 6.8 standard deviations. Structure `B' is observed in the right ascension range of $110^{\circ}$ to $140^{\circ}$. The relative excess observed in this region is at the level of $(4.9\pm1.4)\times10^{-4}$ with a statistical significance of 4.7 standard deviations. These structures are consistent with those reported by Milagro, ARGO-YBJ, and HAWC. These observations could provide a better understanding of the cosmic ray sources, propagation and the magnetic structures in our Galaxy.
△ Less
Submitted 23 October, 2023;
originally announced October 2023.
-
ELVIS: Empowering Locality of Vision Language Pre-training with Intra-modal Similarity
Authors:
Sumin Seo,
JaeWoong Shin,
Jaewoo Kang,
Tae Soo Kim,
Thijs Kooi
Abstract:
Deep learning has shown great potential in assisting radiologists in reading chest X-ray (CXR) images, but its need for expensive annotations for improving performance prevents widespread clinical application. Visual language pre-training (VLP) can alleviate the burden and cost of annotation by leveraging routinely generated reports for radiographs, which exist in large quantities as well as in pa…
▽ More
Deep learning has shown great potential in assisting radiologists in reading chest X-ray (CXR) images, but its need for expensive annotations for improving performance prevents widespread clinical application. Visual language pre-training (VLP) can alleviate the burden and cost of annotation by leveraging routinely generated reports for radiographs, which exist in large quantities as well as in paired form (image-text pairs). Additionally, extensions to localization-aware VLPs are being proposed to address the needs for accurate localization of abnormalities for computer-aided diagnosis (CAD) in CXR. However, we find that the formulation proposed by locality-aware VLP literature actually leads to a loss in spatial relationships required for downstream localization tasks. Therefore, we propose Empowering Locality of VLP with Intra-modal Similarity, ELVIS, a VLP aware of intra-modal locality, to better preserve the locality within radiographs or reports, which enhances the ability to comprehend location references in text reports. Our locality-aware VLP method significantly outperforms state-of-the art baselines in multiple segmentation tasks and the MS-CXR phrase grounding task. Qualitatively, we show that ELVIS focuses well on regions of interest described in the report text compared to prior approaches, allowing for enhanced interpretability.
△ Less
Submitted 23 July, 2023; v1 submitted 11 April, 2023;
originally announced April 2023.
-
Enhancing Breast Cancer Risk Prediction by Incorporating Prior Images
Authors:
Hyeonsoo Lee,
Junha Kim,
Eunkyung Park,
Minjeong Kim,
Taesoo Kim,
Thijs Kooi
Abstract:
Recently, deep learning models have shown the potential to predict breast cancer risk and enable targeted screening strategies, but current models do not consider the change in the breast over time. In this paper, we present a new method, PRIME+, for breast cancer risk prediction that leverages prior mammograms using a transformer decoder, outperforming a state-of-the-art risk prediction method th…
▽ More
Recently, deep learning models have shown the potential to predict breast cancer risk and enable targeted screening strategies, but current models do not consider the change in the breast over time. In this paper, we present a new method, PRIME+, for breast cancer risk prediction that leverages prior mammograms using a transformer decoder, outperforming a state-of-the-art risk prediction method that only uses mammograms from a single time point. We validate our approach on a dataset with 16,113 exams and further demonstrate that it effectively captures patterns of changes from prior mammograms, such as changes in breast density, resulting in improved short-term and long-term breast cancer risk prediction. Experimental results show that our model achieves a statistically significant improvement in performance over the state-of-the-art based model, with a C-index increase from 0.68 to 0.73 (p < 0.05) on held-out test sets.
△ Less
Submitted 28 August, 2023; v1 submitted 27 March, 2023;
originally announced March 2023.
-
Understanding metric-related pitfalls in image analysis validation
Authors:
Annika Reinke,
Minu D. Tizabi,
Michael Baumgartner,
Matthias Eisenmann,
Doreen Heckmann-Nötzel,
A. Emre Kavur,
Tim Rädsch,
Carole H. Sudre,
Laura Acion,
Michela Antonelli,
Tal Arbel,
Spyridon Bakas,
Arriel Benis,
Matthew Blaschko,
Florian Buettner,
M. Jorge Cardoso,
Veronika Cheplygina,
Jianxu Chen,
Evangelia Christodoulou,
Beth A. Cimini,
Gary S. Collins,
Keyvan Farahani,
Luciana Ferrer,
Adrian Galdran,
Bram van Ginneken
, et al. (53 additional authors not shown)
Abstract:
Validation metrics are key for the reliable tracking of scientific progress and for bridging the current chasm between artificial intelligence (AI) research and its translation into practice. However, increasing evidence shows that particularly in image analysis, metrics are often chosen inadequately in relation to the underlying research problem. This could be attributed to a lack of accessibilit…
▽ More
Validation metrics are key for the reliable tracking of scientific progress and for bridging the current chasm between artificial intelligence (AI) research and its translation into practice. However, increasing evidence shows that particularly in image analysis, metrics are often chosen inadequately in relation to the underlying research problem. This could be attributed to a lack of accessibility of metric-related knowledge: While taking into account the individual strengths, weaknesses, and limitations of validation metrics is a critical prerequisite to making educated choices, the relevant knowledge is currently scattered and poorly accessible to individual researchers. Based on a multi-stage Delphi process conducted by a multidisciplinary expert consortium as well as extensive community feedback, the present work provides the first reliable and comprehensive common point of access to information on pitfalls related to validation metrics in image analysis. Focusing on biomedical image analysis but with the potential of transfer to other fields, the addressed pitfalls generalize across application domains and are categorized according to a newly created, domain-agnostic taxonomy. To facilitate comprehension, illustrations and specific examples accompany each pitfall. As a structured body of information accessible to researchers of all levels of expertise, this work enhances global comprehension of a key topic in image analysis validation.
△ Less
Submitted 23 February, 2024; v1 submitted 3 February, 2023;
originally announced February 2023.
-
OOOE: Only-One-Object-Exists Assumption to Find Very Small Objects in Chest Radiographs
Authors:
Gunhee Nam,
Taesoo Kim,
Sanghyup Lee,
Thijs Kooi
Abstract:
The accurate localization of inserted medical tubes and parts of human anatomy is a common problem when analyzing chest radiographs and something deep neural networks could potentially automate. However, many foreign objects like tubes and various anatomical structures are small in comparison to the entire chest X-ray, which leads to severely unbalanced data and makes training deep neural networks…
▽ More
The accurate localization of inserted medical tubes and parts of human anatomy is a common problem when analyzing chest radiographs and something deep neural networks could potentially automate. However, many foreign objects like tubes and various anatomical structures are small in comparison to the entire chest X-ray, which leads to severely unbalanced data and makes training deep neural networks difficult. In this paper, we present a simple yet effective `Only-One-Object-Exists' (OOOE) assumption to improve the deep network's ability to localize small landmarks in chest radiographs. The OOOE enables us to recast the localization problem as a classification problem and we can replace commonly used continuous regression techniques with a multi-class discrete objective. We validate our approach using a large scale proprietary dataset of over 100K radiographs as well as publicly available RANZCR-CLiP Kaggle Challenge dataset and show that our method consistently outperforms commonly used regression-based detection models as well as commonly used pixel-wise classification methods. Additionally, we find that the method using the OOOE assumption generalizes to multiple detection problems in chest X-rays and the resulting model shows state-of-the-art performance on detecting various tube tips inserted to the patient as well as patient anatomy.
△ Less
Submitted 13 October, 2022;
originally announced October 2022.
-
Did You Get What You Paid For? Rethinking Annotation Cost of Deep Learning Based Computer Aided Detection in Chest Radiographs
Authors:
Tae Soo Kim,
Geonwoon Jang,
Sanghyup Lee,
Thijs Kooi
Abstract:
As deep networks require large amounts of accurately labeled training data, a strategy to collect sufficiently large and accurate annotations is as important as innovations in recognition methods. This is especially true for building Computer Aided Detection (CAD) systems for chest X-rays where domain expertise of radiologists is required to annotate the presence and location of abnormalities on X…
▽ More
As deep networks require large amounts of accurately labeled training data, a strategy to collect sufficiently large and accurate annotations is as important as innovations in recognition methods. This is especially true for building Computer Aided Detection (CAD) systems for chest X-rays where domain expertise of radiologists is required to annotate the presence and location of abnormalities on X-ray images. However, there lacks concrete evidence that provides guidance on how much resource to allocate for data annotation such that the resulting CAD system reaches desired performance. Without this knowledge, practitioners often fall back to the strategy of collecting as much detail as possible on as much data as possible which is cost inefficient. In this work, we investigate how the cost of data annotation ultimately impacts the CAD model performance on classification and segmentation of chest abnormalities in frontal-view X-ray images. We define the cost of annotation with respect to the following three dimensions: quantity, quality and granularity of labels. Throughout this study, we isolate the impact of each dimension on the resulting CAD model performance on detecting 10 chest abnormalities in X-rays. On a large scale training data with over 120K X-ray images with gold-standard annotations, we find that cost-efficient annotations provide great value when collected in large amounts and lead to competitive performance when compared to models trained with only gold-standard annotations. We also find that combining large amounts of cost efficient annotations with only small amounts of expensive labels leads to competitive CAD models at a much lower cost.
△ Less
Submitted 30 September, 2022;
originally announced September 2022.
-
Tutorial on the development of AI models for medical image analysis
Authors:
Thijs Kooi
Abstract:
The idea of using computers to read medical scans was introduced as early as 1966. However, limits to machine learning technology meant progress was slow initially. The Alexnet breakthrough in 2012 sparked new interest in the topic, which resulted in the release of 100s of medical AI solutions on the market. In spite of success for some diseases and modalities, many challenges remain. Research typ…
▽ More
The idea of using computers to read medical scans was introduced as early as 1966. However, limits to machine learning technology meant progress was slow initially. The Alexnet breakthrough in 2012 sparked new interest in the topic, which resulted in the release of 100s of medical AI solutions on the market. In spite of success for some diseases and modalities, many challenges remain. Research typically focuses on the development of specific applications or techniques, clinical evaluation, or meta analysis of clinical studies or techniques through surveys or challenges. However, limited attention has been given to the development process of improving real world performance. In this tutorial, we address the latter and discuss some techniques to conduct the development process in order to make this as efficient as possible.
△ Less
Submitted 14 July, 2022;
originally announced August 2022.
-
Proton Penetration Efficiency over a High Altitude Observatory in Mexico
Authors:
S. Miyake,
T. Koi,
Y. Muraki,
Y. Matsubara,
S. Masuda,
P. Miranda,
T. Naito,
E. Ortiz,
A. Oshima,
T. Sakai,
T. Sako,
S. Shibata,
H. Takamaru,
M. Tokumaru,
J. F. Valdes-Galicia
Abstract:
In association with a large solar flare on November 7, 2004, the solar neutron detectors located at Mt. Chacaltaya (5,250m) in Bolivia and Mt. Sierra Negra (4,600m) in Mexico recorded very interesting events. In order to explain these events, we have performed a calculation solving the equation of motion of anti-protons inside the magnetosphere. Based on these results, the Mt. Chacaltaya event may…
▽ More
In association with a large solar flare on November 7, 2004, the solar neutron detectors located at Mt. Chacaltaya (5,250m) in Bolivia and Mt. Sierra Negra (4,600m) in Mexico recorded very interesting events. In order to explain these events, we have performed a calculation solving the equation of motion of anti-protons inside the magnetosphere. Based on these results, the Mt. Chacaltaya event may be explained by the detection of solar neutrons, while the Mt. Sierra Negra event may be explained by the first detection of very high energy solar neutron decay protons (SNDPs) around 6 GeV.
△ Less
Submitted 8 July, 2022; v1 submitted 5 July, 2022;
originally announced July 2022.
-
Detectability of southern gamma-ray sources beyond 100 TeV with ALPAQUITA, the prototype experiment of ALPACA
Authors:
S. Kato,
C. A. H. Condori,
E. dela Fuente,
A. Gomi,
K. Hibino,
N. Hotta,
I. Toledano-Juarez,
Y. Katayose,
C. Kato,
K. Kawata,
W. Kihara,
Y. Ko,
T. Koi,
H. Kojima,
D. Kurashige,
J. Lozoya,
F. Orozco-Luna,
R. Mayta,
P. Miranda,
K. Munakata,
H. Nakada,
Y. Nakamura,
Y. Nakazawa,
C. Nina,
M. Nishizawa
, et al. (25 additional authors not shown)
Abstract:
Andes Large-area PArticle detector for Cosmic-ray physics and Astronomy (ALPACA) is an international experiment that applies southern very-high-energy (VHE) gamma-ray astronomy to determine the origin of cosmic rays around the knee energy region ($10^{15}\, {\rm eV} - 10^{16}\, {\rm eV}$). The experiment consists of an air shower (AS) array with a surface of $83, 000\, {\rm m}^2$ and an undergroun…
▽ More
Andes Large-area PArticle detector for Cosmic-ray physics and Astronomy (ALPACA) is an international experiment that applies southern very-high-energy (VHE) gamma-ray astronomy to determine the origin of cosmic rays around the knee energy region ($10^{15}\, {\rm eV} - 10^{16}\, {\rm eV}$). The experiment consists of an air shower (AS) array with a surface of $83, 000\, {\rm m}^2$ and an underground water Cherenkov muon detector (MD) array covering $5, 400\, {\rm m}^2$. The experimental site is at the Mt. Chacaltaya plateau in La Paz, Bolivia, with an altitude of $4, 740\, {\rm m}$ corresponding to $572\, {\rm g}/{\rm cm}^2$ atmospheric thickness. As the prototype experiment of ALPACA, the ALPAQUITA experiment aims to begin data acquisition in late 2021. The ALPAQUITA array consists of a smaller AS array ($18, 450\, {\rm m}^2$) and underground MD ($900\, {\rm m}^2$), which are now under construction. ALPAQUITA's sensitivity to gamma-ray sources is evaluated with Monte Carlo simulations. The simulation finds that five gamma-ray sources observed by H.E.S.S. and HAWC experiments will be detected by ALPAQUITA beyond 10 TeV and one out of these five - HESS J1702-420A - above 300 TeV in one calendar year observation. The latter finding means that scientific discussions can be made on the emission mechanism of gamma rays beyond 100 TeV from southern sources on the basis of the observational results of this prototype experiment.
△ Less
Submitted 28 September, 2021;
originally announced September 2021.
-
Common Limitations of Image Processing Metrics: A Picture Story
Authors:
Annika Reinke,
Minu D. Tizabi,
Carole H. Sudre,
Matthias Eisenmann,
Tim Rädsch,
Michael Baumgartner,
Laura Acion,
Michela Antonelli,
Tal Arbel,
Spyridon Bakas,
Peter Bankhead,
Arriel Benis,
Matthew Blaschko,
Florian Buettner,
M. Jorge Cardoso,
Jianxu Chen,
Veronika Cheplygina,
Evangelia Christodoulou,
Beth Cimini,
Gary S. Collins,
Sandy Engelhardt,
Keyvan Farahani,
Luciana Ferrer,
Adrian Galdran,
Bram van Ginneken
, et al. (68 additional authors not shown)
Abstract:
While the importance of automatic image analysis is continuously increasing, recent meta-research revealed major flaws with respect to algorithm validation. Performance metrics are particularly key for meaningful, objective, and transparent performance assessment and validation of the used automatic algorithms, but relatively little attention has been given to the practical pitfalls when using spe…
▽ More
While the importance of automatic image analysis is continuously increasing, recent meta-research revealed major flaws with respect to algorithm validation. Performance metrics are particularly key for meaningful, objective, and transparent performance assessment and validation of the used automatic algorithms, but relatively little attention has been given to the practical pitfalls when using specific metrics for a given image analysis task. These are typically related to (1) the disregard of inherent metric properties, such as the behaviour in the presence of class imbalance or small target structures, (2) the disregard of inherent data set properties, such as the non-independence of the test cases, and (3) the disregard of the actual biomedical domain interest that the metrics should reflect. This living dynamically document has the purpose to illustrate important limitations of performance metrics commonly applied in the field of image analysis. In this context, it focuses on biomedical image analysis problems that can be phrased as image-level classification, semantic segmentation, instance segmentation, or object detection task. The current version is based on a Delphi process on metrics conducted by an international consortium of image analysis experts from more than 60 institutions worldwide.
△ Less
Submitted 6 December, 2023; v1 submitted 12 April, 2021;
originally announced April 2021.
-
Solar Neutron Decay Protons observed in November 7, 2004
Authors:
Yasushi Muraki,
Jose F. Valde-Galicia,
Ernesto Ortiz,
Yutaka Matsubara,
Shoichi Shibata,
Takashi Sako,
Satoshi Masuda,
Munetoshi Tokumaru,
Tatsumi Koi,
Akitoshi Ooshima,
Takasuke Sakai,
Tsuguya Naito,
Pedro Miranda
Abstract:
We have found an interesting event registered by the solar neutron telescopes installed at high mountains in Bolivia (5250 m a.s.l.) and Mexico (4600 m a.s.l.). The event was observed November 7th of 2004 in association with a large solar flare of magnitude X2.0. Some features in our registers and in two satellites (GOES 11 and SOHO) reveal the presence of electrons and protons as possible product…
▽ More
We have found an interesting event registered by the solar neutron telescopes installed at high mountains in Bolivia (5250 m a.s.l.) and Mexico (4600 m a.s.l.). The event was observed November 7th of 2004 in association with a large solar flare of magnitude X2.0. Some features in our registers and in two satellites (GOES 11 and SOHO) reveal the presence of electrons and protons as possible products of neutron decay. Solar neutron decay protons (sndp) were recorded on board ISEE3 satellite in June 3rd, 1982 . On October 19th, 1989, the ground level detectors installed in Goose Bay and Deep River revealed the registration of solar neutron decay protons (sndp). Therefore this is the second example that such an evidence is registered on the Earth's surface.
△ Less
Submitted 31 December, 2020;
originally announced December 2020.
-
Origin of the in-orbit instrumental background of the Hard X-ray Imager onboard Hitomi
Authors:
Kouichi Hagino,
Hirokazu Odaka,
Goro Sato,
Tamotsu Sato,
Hiromasa Suzuki,
Tsunefumi Mizuno,
Madoka Kawaharada,
Masanori Ohno,
Kazuhiro Nakazawa,
Shogo B. Kobayashi,
Hiroaki Murakami,
Katsuma Miyake,
Makoto Asai,
Tatsumi Koi,
Greg Madejski,
Shinya Saito,
Dennis H. Wright,
Teruaki Enoto,
Yasushi Fukazawa,
Katsuhiro Hayashi,
Jun Kataoka,
Junichiro Katsuta,
Motohide Kokubun,
Philippe Laurent,
Francois Lebrun
, et al. (21 additional authors not shown)
Abstract:
Understanding and reducing the in-orbit instrumental backgrounds are essential to achieving high sensitivity in hard X-ray astronomical observations. The observational data of the Hard X-ray Imager (HXI) on board the Hitomi satellite provides useful information on the background components, owing to its multi-layer configuration with different atomic numbers: the HXI consists of a stack of four la…
▽ More
Understanding and reducing the in-orbit instrumental backgrounds are essential to achieving high sensitivity in hard X-ray astronomical observations. The observational data of the Hard X-ray Imager (HXI) on board the Hitomi satellite provides useful information on the background components, owing to its multi-layer configuration with different atomic numbers: the HXI consists of a stack of four layers of Si (Z = 14) detectors and one layer of CdTe (Z = 48, 52) detector surrounded by well-type BGO (Bi4Ge3O12) active shields. Based on the observational data, the backgrounds of top Si layer, the three underlying Si layers, and the CdTe layer are inferred to be dominated by different components, namely, low-energy electrons, albedo neutrons, and proton-induced radioactivation, respectively. Monte Carlo simulations of the in-orbit background of the HXI reproduce the observed background spectrum of each layer well, thereby verifying the above hypothesis quantitatively. In addition, we suggest the inclusion of an electron shield to reduce the background.
△ Less
Submitted 20 November, 2020;
originally announced November 2020.
-
GEANT4 Parameter Tuning Using Professor
Authors:
V. Elvira,
L. Fields,
K. L. Genser,
R. Hatcher,
V. Ivanchenko,
M. Kelsey,
T. Koi,
G. N. Perdue,
A. Ribon,
V. Uzhinsky,
D. H. Wright,
J. Yarba,
S. Y. Jun
Abstract:
The Geant4 toolkit is used extensively in high energy physics to simulate the passage of particles through matter and to predict effects such as detector efficiencies and smearing. Geant4 uses many underlying models to predict particle interaction kinematics, and uncertainty in these models leads to uncertainty in high energy physics measurements. The Geant4 collaboration recently made free parame…
▽ More
The Geant4 toolkit is used extensively in high energy physics to simulate the passage of particles through matter and to predict effects such as detector efficiencies and smearing. Geant4 uses many underlying models to predict particle interaction kinematics, and uncertainty in these models leads to uncertainty in high energy physics measurements. The Geant4 collaboration recently made free parameters in some models accessible through partnership with Geant4 developers. We present a study of the impact of varying parameters in three Geant4 hadronic physics models on agreement with thin target datasets and describe fits to these datasets using the Professor model tuning framework. We find that varying parameters produces substantially better agreement with some datasets, but that more degrees of freedom are required for full agreement. This work is a first step towards a common framework for propagating uncertainties in Geant4 models to high energy physics measurements, and we outline future work required to complete that goal.
△ Less
Submitted 16 June, 2020; v1 submitted 14 October, 2019;
originally announced October 2019.
-
Error Exponents for Asynchronous Multiple Access Channels. Controlled Asynchronism may Outperform Synchronism
Authors:
Lóránt Farkas,
Tamás Kói
Abstract:
Exponential error bounds achievable by universal coding and decoding are derived for frame-asynchronous discrete memoryless %asynchronous multiple access channels with two senders, via the method of subtypes, a refinement of the method of types. Maximum empirical multi-information decoding is employed. A key tool is an improved packing lemma, that overcomes the technical difficulty caused by codew…
▽ More
Exponential error bounds achievable by universal coding and decoding are derived for frame-asynchronous discrete memoryless %asynchronous multiple access channels with two senders, via the method of subtypes, a refinement of the method of types. Maximum empirical multi-information decoding is employed. A key tool is an improved packing lemma, that overcomes the technical difficulty caused by codeword repetitions, via an induction based new argument. The asymptotic form of the bounds admits numerical evaluation. This demostrates that error exponents achievable by synchronous transmission (if possible) can be superseeded via controlled asynchronism, i.e. a deliberate shift of the codewords.
△ Less
Submitted 3 February, 2020; v1 submitted 11 July, 2019;
originally announced July 2019.
-
Po-production in lead: A benchmark between Geant4, FLUKA and MCNPX
Authors:
Alfredo Ferrari,
Daniela Kiselev,
Tatsumi Koi,
Michael Wohlmuther,
Jean-Christophe Davide
Abstract:
On the last SATIF a comparison between the measured activities of the polonium isotopes Po-208, Po-209 and Po-210 and the simulated results using MCNPX2.7.0 was presented. The lead samples were cut from the SINQ spallation target at the Paul Scherrer Institut (PSI) and irradiated in 2000/2001 by 575 MeV protons. The Po-isotopes were separated using radiochemical methods by the group of D. Schumann…
▽ More
On the last SATIF a comparison between the measured activities of the polonium isotopes Po-208, Po-209 and Po-210 and the simulated results using MCNPX2.7.0 was presented. The lead samples were cut from the SINQ spallation target at the Paul Scherrer Institut (PSI) and irradiated in 2000/2001 by 575 MeV protons. The Po-isotopes were separated using radiochemical methods by the group of D. Schumann at PSI and measured. Choosing the default model in MCNPX, Bertini-Dresner, the prediction underestimated the measured activities by up to several orders of magnitude. Therefore the Liège intranuclear-cascade model (INCL4.6) coupled to the de-excitation model ABLA07 were implemented into MCNPX2.7.0 and very good agreement was found to the measurement. The reason for the disagreement was traced back to a suppression of alpha reactions on the lead isotopes leading to Po and neglecting the triton capture on Pb-208, which leads to Pb-210 and decays into Po-210 with a much longer life time (22.3 years) than the decay of Po-210 itself (138 days). The prediction of the Po-isotope activities turns out to be a sensitive test for models and codes as it requires the accurate treatment of reaction channels not only with neutrons, protons and pions but also with alphas and tritons, which are not considered in intra-nuclear cascade models of the first generation. Therefore it was decided to perform a benchmark by comparing the results obtained with MCNPX2.7.0 using INCL4.6/ABLA07 to the predictions of FLUKA and Geant4. Since the model of the SINQ spallation source requires an elaborate geometry a toy model was setup. The toy model has a simplified geometry preserving the main features of the original geometry. The results for the activities of the three Po-isotopes and Pb-210 as well as the energy spectra for alphas and tritons obtained with the three particle transport Monte Carlo codes are presented.
△ Less
Submitted 10 June, 2018;
originally announced June 2018.
-
Modeling of proton-induced radioactivation background in hard X-ray telescopes: Geant4-based simulation and its demonstration by Hitomi's measurement in a low Earth orbit
Authors:
Hirokazu Odaka,
Makoto Asai,
Kouichi Hagino,
Tatsumi Koi,
Greg Madejski,
Tsunefumi Mizuno,
Masanori Ohno,
Shinya Saito,
Tamotsu Sato,
Dennis H. Wright,
Teruaki Enoto,
Yasushi Fukazawa,
Katsuhiro Hayashi,
Jun Kataoka,
Junichiro Katsuta,
Madoka Kawaharada,
Shogo B. Kobayashi,
Motohide Kokubun,
Philippe Laurent,
Francois Lebrun,
Olivier Limousin,
Daniel Maier,
Kazuo Makishima,
Taketo Mimura,
Katsuma Miyake
, et al. (25 additional authors not shown)
Abstract:
Hard X-ray astronomical observatories in orbit suffer from a significant amount of background due to radioactivation induced by cosmic-ray protons and/or geomagnetically trapped protons. Within the framework of a full Monte Carlo simulation, we present modeling of in-orbit instrumental background dominated by radioactivation. To reduce the computation time required by straightforward simulations o…
▽ More
Hard X-ray astronomical observatories in orbit suffer from a significant amount of background due to radioactivation induced by cosmic-ray protons and/or geomagnetically trapped protons. Within the framework of a full Monte Carlo simulation, we present modeling of in-orbit instrumental background dominated by radioactivation. To reduce the computation time required by straightforward simulations of delayed emissions from activated isotopes, we insert a semi-analytical calculation that converts production probabilities of radioactive isotopes by interaction of the primary protons into decay rates at measurement time of all secondary isotopes. Therefore, our simulation method is separated into three steps: (1) simulation of isotope production, (2) semi-analytical conversion to decay rates, and (3) simulation of decays of the isotopes at measurement time. This method is verified by a simple setup that has a CdTe semiconductor detector, and shows a 100-fold improvement in efficiency over the straightforward simulation. The simulation framework was tested against data measured with a CdTe sensor in the Hard X-ray Imager onboard the Hitomi X-ray Astronomy Satellite, which was put into a low Earth orbit with an altitude of 570 km and an inclination of 31 degrees, and thus experienced a large amount of irradiation from geomagnetically trapped protons during its passages through the South Atlantic Anomaly. The simulation is able to treat full histories of the proton irradiation and multiple measurement windows. The simulation results agree very well with the measured data, showing that the measured background is well described by the combination of proton-induced radioactivation of the CdTe detector itself and thick Bi4Ge3O12 scintillator shields, leakage of cosmic X-ray background and albedo gamma-ray radiation, and emissions from naturally contaminated isotopes in the detector system.
△ Less
Submitted 3 April, 2018;
originally announced April 2018.
-
Classifying Symmetrical Differences and Temporal Change in Mammography Using Deep Neural Networks
Authors:
Thijs Kooi,
Nico Karssemeijer
Abstract:
We investigate the addition of symmetry and temporal context information to a deep Convolutional Neural Network (CNN) with the purpose of detecting malignant soft tissue lesions in mammography. We employ a simple linear map** that takes the location of a mass candidate and maps it to either the contra-lateral or prior mammogram and Regions Of Interest (ROI) are extracted around each location. We…
▽ More
We investigate the addition of symmetry and temporal context information to a deep Convolutional Neural Network (CNN) with the purpose of detecting malignant soft tissue lesions in mammography. We employ a simple linear map** that takes the location of a mass candidate and maps it to either the contra-lateral or prior mammogram and Regions Of Interest (ROI) are extracted around each location. We subsequently explore two different architectures (1) a fusion model employing two datastreams were both ROIs are fed to the network during training and testing and (2) a stage-wise approach where a single ROI CNN is trained on the primary image and subsequently used as feature extractor for both primary and symmetrical or prior ROIs. A 'shallow' Gradient Boosted Tree (GBT) classifier is then trained on the concatenation of these features and used to classify the joint representation. Results shown a significant increase in performance using the first architecture and symmetry information, but only marginal gains in performance using temporal data and the other setting. We feel results are promising and can greatly be improved when more temporal data becomes available.
△ Less
Submitted 1 August, 2017; v1 submitted 22 March, 2017;
originally announced March 2017.
-
A Survey on Deep Learning in Medical Image Analysis
Authors:
Geert Litjens,
Thijs Kooi,
Babak Ehteshami Bejnordi,
Arnaud Arindra Adiyoso Setio,
Francesco Ciompi,
Mohsen Ghafoorian,
Jeroen A. W. M. van der Laak,
Bram van Ginneken,
Clara I. Sánchez
Abstract:
Deep learning algorithms, in particular convolutional networks, have rapidly become a methodology of choice for analyzing medical images. This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year. We survey the use of deep learning for image classification, object detection, se…
▽ More
Deep learning algorithms, in particular convolutional networks, have rapidly become a methodology of choice for analyzing medical images. This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year. We survey the use of deep learning for image classification, object detection, segmentation, registration, and other tasks and provide concise overviews of studies per application area. Open challenges and directions for future research are discussed.
△ Less
Submitted 4 June, 2017; v1 submitted 19 February, 2017;
originally announced February 2017.
-
Universal Random Access Error Exponents for Codebooks of Different Word-Lengths
Authors:
Lóránt Farkas,
Tamás Kói
Abstract:
Csiszár's channel coding theorem for multiple codebooks is generalized allowing the codeword lenghts differ across codebooks. Also in this case, for each codebook an error exponent can be achieved that equals the random coding exponent for this codebook alone, in addition, erasure detection failure probability tends to 0. This is proved even for sender and receiver not knowing the channel. As a co…
▽ More
Csiszár's channel coding theorem for multiple codebooks is generalized allowing the codeword lenghts differ across codebooks. Also in this case, for each codebook an error exponent can be achieved that equals the random coding exponent for this codebook alone, in addition, erasure detection failure probability tends to 0. This is proved even for sender and receiver not knowing the channel. As a corollary, a substantial improvement is obtained when the sender knows the channel.
△ Less
Submitted 22 January, 2017; v1 submitted 7 July, 2016;
originally announced July 2016.
-
Random Access and Source-Channel Coding Error Exponents for Multiple Access Channels
Authors:
Lóránt Farkas,
Tamás Kói
Abstract:
A new universal coding/decoding scheme for random access with collision detection is given in the case of two senders. The result is used to give an achievable joint source-channel coding error exponent for multiple access channels in the case of independent sources. This exponent is improved in a modified model that admits error free 0 rate communication between the senders.
A new universal coding/decoding scheme for random access with collision detection is given in the case of two senders. The result is used to give an achievable joint source-channel coding error exponent for multiple access channels in the case of independent sources. This exponent is improved in a modified model that admits error free 0 rate communication between the senders.
△ Less
Submitted 18 September, 2013; v1 submitted 27 January, 2013;
originally announced January 2013.
-
On Capacity Regions of Discrete Asynchronous Multiple Access Channels
Authors:
Lóránt Farkas,
Tamás Kói
Abstract:
A general formalization is given for asynchronous multiple access channels which admits different assumptions on delays. This general framework allows the analysis of so far unexplored models leading to new interesting capacity regions. In particular, a single letter characterization is given for the capacity region in case of 3 senders, 2 synchronous with each other and the third not synchronous…
▽ More
A general formalization is given for asynchronous multiple access channels which admits different assumptions on delays. This general framework allows the analysis of so far unexplored models leading to new interesting capacity regions. In particular, a single letter characterization is given for the capacity region in case of 3 senders, 2 synchronous with each other and the third not synchronous with them.
△ Less
Submitted 12 July, 2014; v1 submitted 11 April, 2012;
originally announced April 2012.
-
Testability of minimum balanced multiway cut densities
Authors:
Marianna Bolla,
Tamas Koi,
Andras Kramli
Abstract:
Testable weighted graph parameters and equivalent notions of testability are investigated based on papers of Laszlo Lovasz and coauthors. We prove that certain balanced minimum multiway cut densities are testable. Using this fact, quadratic programming techniques are applied to approximate some of these quantities. The problem is related to cluster analysis and statistical physics. Convergence o…
▽ More
Testable weighted graph parameters and equivalent notions of testability are investigated based on papers of Laszlo Lovasz and coauthors. We prove that certain balanced minimum multiway cut densities are testable. Using this fact, quadratic programming techniques are applied to approximate some of these quantities. The problem is related to cluster analysis and statistical physics. Convergence of special noisy graph sequences is also discussed.
△ Less
Submitted 11 January, 2010;
originally announced January 2010.
-
Measurements of total and partial charge-changing cross sections for 200-400 MeV/nucleon 12C in water and polycarbonate
Authors:
T. Toshito,
K. Kodama,
L. Sihver,
K. Yusa,
M. Ozaki,
K. Amako,
S. Kameoka,
K. Murakami,
T. Sasaki,
S. Aoki,
T. Ban,
T. Fukuda,
M. Komatsu,
H. Kubota,
N. Naganawa,
T. Nakamura,
T. Nakano,
M. Natsume,
K. Niwa,
S. Takahashi,
J. Yoshida,
H. Yoshida,
M. Kanazawa,
N. Kanematsu,
M. Komori
, et al. (7 additional authors not shown)
Abstract:
We have studied charged nuclear fragments produced by 200 - 400 MeV/nucleon carbon ions, interacting with water and polycarbonate, using a newly developed emulsion detector. Total and partial charge-changing cross sections for the production of B, Be, and Li fragments were measured and compared with both previously published measurements, and model predictions. This study is of importance for va…
▽ More
We have studied charged nuclear fragments produced by 200 - 400 MeV/nucleon carbon ions, interacting with water and polycarbonate, using a newly developed emulsion detector. Total and partial charge-changing cross sections for the production of B, Be, and Li fragments were measured and compared with both previously published measurements, and model predictions. This study is of importance for validating and improving carbon ion therapy treatment planning systems, and for estimating the radiological risks for personnel on space missions, since carbon is a significant component of the Galactic Cosmic Rays.
△ Less
Submitted 27 March, 2007;
originally announced March 2007.
-
Parameterization of Gamma, e^+/- and Neutrino Spectra Produced by p-p Interaction in Astronomical Environment
Authors:
Tuneyoshi Kamae,
Niklas Karlsson,
Tsunefumi Mizuno,
Toshinori Abe,
Tatsumi Koi
Abstract:
We present the yield and spectra of stable secondary particles (gamma, e^+/-, nu_e, nubar_e, nu_mu, and nubar_mu) of p-p interaction in parameterized formulae to facilitate calculations involving them in astronomical environments. The formulae are derived on the up-to-date p-p interaction model by [Kamae05] which incorporates the logarithmically rising inelastic cross-section, the diffraction di…
▽ More
We present the yield and spectra of stable secondary particles (gamma, e^+/-, nu_e, nubar_e, nu_mu, and nubar_mu) of p-p interaction in parameterized formulae to facilitate calculations involving them in astronomical environments. The formulae are derived on the up-to-date p-p interaction model by [Kamae05] which incorporates the logarithmically rising inelastic cross-section, the diffraction dissociation process, and the Feynman scaling violation. To improve fidelity to experimental data in lower energies, two baryon resonance contributions have been added: one representing Delta(1232) and the other multiple resonances around 1600MeV/c^2. The parametrized formulae predict that all secondary particle spectra be harder by about 0.05 in power-law indices than that of the incident proton and their inclusive cross-sections be larger than those predicted by p-p interaction models based on the Feynman scaling.
△ Less
Submitted 16 January, 2007; v1 submitted 23 May, 2006;
originally announced May 2006.
-
Diffractive Interaction and Scaling Violation in pp->pi^0 Interaction and GeV Excess in Galactic Diffuse Gamma-Ray Spectrum of EGRET
Authors:
T. Kamae,
T. Abe,
T. Koi
Abstract:
We present here a new calculation of the gamma-ray spectrum from pp->pi^0 in the Galactic ridge environment. The calculation includes the diffractive pp interaction and incorporates the Feynman scaling violation for the first time. Galactic diffuse gamma-rays come, predominantly, from pi^0->gamma gamma in the sub-GeV to multi-GeV range. Hunter et al. found, however, an excess in the GeV range ("…
▽ More
We present here a new calculation of the gamma-ray spectrum from pp->pi^0 in the Galactic ridge environment. The calculation includes the diffractive pp interaction and incorporates the Feynman scaling violation for the first time. Galactic diffuse gamma-rays come, predominantly, from pi^0->gamma gamma in the sub-GeV to multi-GeV range. Hunter et al. found, however, an excess in the GeV range ("GeV Excess") in the EGRET Galactic diffuse spectrum above the prediction based on experimental pp->pi^0 cross-sections and the Feynman scaling hypothesis. We show, in this work, that the diffractive process makes the gamma-ray spectrum harder than the incident proton spectrum by ~0.05 in power-law index, and, that the scaling violation produces 30-80% more pi^0 than the scaling model for incident proton energies above 100GeV. Combination of the two can explain about a half of the "GeV Excess" with the local cosmic proton (power-law index ~2.7). The excess can be fully explained if the proton spectral index in the Galactic ridge is a little harder (~0.2 in power-law index) than the local spectrum. Given also in the paper is that the diffractive process enhances e^+ over e^- and the scaling violation gives 50-100% higher p-bar yield than without the violation, both in the multi-GeV range.
△ Less
Submitted 21 January, 2005; v1 submitted 26 October, 2004;
originally announced October 2004.
-
Interfacing the JQMD and JAM Nuclear Reaction Codes to Geant4
Authors:
T. Koi,
M. Asai,
D. H. Wright,
K. Niita,
Y. Nara,
K. Amako,
T. Sasaki
Abstract:
Geant4 is a toolkit for the simulation of the passage of particles through matter. It provides a comprehensive set of tools for geometry, tracking, detector response, run, event and track management, visualization and user interfaces. Geant4 also has an abundant set of physics models that handle the diverse interactions of particles with matter across a wide energy range. However, there are also…
▽ More
Geant4 is a toolkit for the simulation of the passage of particles through matter. It provides a comprehensive set of tools for geometry, tracking, detector response, run, event and track management, visualization and user interfaces. Geant4 also has an abundant set of physics models that handle the diverse interactions of particles with matter across a wide energy range. However, there are also many well-established reaction codes currently used in the same fields where Geant4 is applied. In order to take advantage of these codes, we began to investigate their use from within the framework of Geant4. The first codes chosen for this investigation were the Jaeri Quantum Molecular Dynamics (JQMD) and Jet AA Microscopic Transportation Model (JAM) codes. JQMD is a QMD model code which is widely used to analyze various aspects of heavy ion reactions. JAM is a hadronic cascade model code which explicitly treats all established hadronic states, including resonances with explicit spin and isospin, as well as their anti-particles. We successfully developed interfaces between these codes and Geant4. These allow a user to construct a detector using the powerful material and geometrical capabilities of Geant4, while at the same time implementing nuclear reactions handled by the JQMD and JAM models and the Hadronic framework of Geant4 proved its flexibility and expandability.
△ Less
Submitted 13 June, 2003;
originally announced June 2003.
-
Development of an Interface for Using EGS4 Physics Processes in Geant4
Authors:
K. Murakami,
K. Amako,
H. Hirayama,
Y. Namito,
T. Sasaki,
M. Asai,
T. Koi
Abstract:
As simulation system, the variety of physics processes implemented is one of the most important functionalities. In that sense, Geant4 is one of the most powerful simulation toolkits. Its flexibility and expansibility brought by object-oriented approach make it possible for us to easily assimilate external simulation packages into the Geant4 system as modules of physics processes. We developed a…
▽ More
As simulation system, the variety of physics processes implemented is one of the most important functionalities. In that sense, Geant4 is one of the most powerful simulation toolkits. Its flexibility and expansibility brought by object-oriented approach make it possible for us to easily assimilate external simulation packages into the Geant4 system as modules of physics processes. We developed an interface for using EGS4, which is another of the most well-known simulation package for electromagnetic physics, in Geant4. By means of this interface, EGS4 users can share Geant4 powerful resources, such as geometry, tracking, etc. It is also important that it can provide a common environment for comparison tests between EGS4 and Geant4. In this paper, we describe our design and implementation of the interface.
△ Less
Submitted 10 June, 2003;
originally announced June 2003.
-
The Geant4 Hadronic Verification Suite for the Cascade Energy Range
Authors:
V. Ivanchenko,
G. Folger,
J. P. Wellisch,
T. Koi,
D. H. Wright
Abstract:
A Geant4 hadronic process verification suite has been designed to test and optimize Geant4 hadronic models in the cascade energy range. It focuses on quantities relevant to the LHC radiation environment and spallation source targets. The general structure of the suite is presented, including the user interface, stages of verification, management of experimental data, event generation, and compar…
▽ More
A Geant4 hadronic process verification suite has been designed to test and optimize Geant4 hadronic models in the cascade energy range. It focuses on quantities relevant to the LHC radiation environment and spallation source targets. The general structure of the suite is presented, including the user interface, stages of verification, management of experimental data, event generation, and comparison of results to data. Verification results for the newly released Binary cascade and Bertini cascade models are presented.
△ Less
Submitted 2 June, 2003;
originally announced June 2003.
