Yaroslav Nikulin - Academia.edu (original) (raw)

Uploads

Papers by Yaroslav Nikulin

A participant's experience of the Digital Mammography DREAM Challenge.

arXiv (Cornell University), May 15, 2016

In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style... more In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style [7]. While showing great results when transferring homogeneous and repetitive patterns, the original style representation often fails to capture more complex properties, like having separate styles of foreground and background. This leads to visual artifacts and undesirable textures appearing in unexpected regions when performing style transfer. We tackle this issue with a variety of approaches, mostly by modifying the style representation in order for it to capture more information and impose a tighter constraint on the style transfer result. In our experiments, we subjectively evaluate our best method as producing from barely noticeable to significant improvements in the quality of style transfer.

Lecture Notes in Computer Science, 2017

As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in th... more As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in their translation to the bed side, in particular for abnormal cardiac rhythm diagnosis and therapy planning. However, the parameters that govern these models need to be estimated from noisy and sparse clinical data in an efficient and precise way, which is still an unsolved challenge. Invasive cardiac mapping provides the richest EP information available today. This paper proposes a new method to estimate a local map of electrical conductivities of the bi-ventricular heart by applying the back-propagation error concept, widely used in neural networks. The method works when either endocardial or epicardial activation time maps are available, and can cope with heterogeneous cardiac tissue. The method was evaluated on synthetic data, showing significantly increased performance in goodness of fit compared to a global parameter estimation approach. The resulting predictive power of the personalized model for cardiac resynchronization therapy was then assessed on 16 swine models of left bundle branch block with rich imaging and EP data before and after CRT. With the proposed personalization, the average error in activation time post CRT was \(10 \pm 4.5\) ms, lower than the observed pre/post-CRT difference of \(26.3 \pm 16.8\) ms.

arXiv (Cornell University), Feb 23, 2016

In this work we explore the method of style transfer presented in [1]. We first demonstrate the p... more In this work we explore the method of style transfer presented in [1]. We first demonstrate the power of the suggested style space on a few examples. We then vary different hyper-parameters and program properties that were not discussed in [1], among which are the recognition network used, starting point of the gradient descent and different ways to partition style and content layers. We also give a brief comparison of some of the existing algorithm implementations and deep learning frameworks used. To study the style space further, an idea similar to [2] is used to generate synthetic images by maximizing a single entry in one of the Gram matrices G l and some interesting results are observed. Next, we try to mimic the sparsity and intensity distribution of Gram matrices obtained from a real painting and generate more complex textures. Finally, we propose two new style representations built on top of network's features and discuss how one could be used to achieve local and potentially content-aware style transfer.

arXiv (Cornell University), Feb 23, 2016

In this work we explore the method of style transfer presented in [1]. We first demonstrate the p... more In this work we explore the method of style transfer presented in [1]. We first demonstrate the power of the suggested style space on a few examples. We then vary different hyper-parameters and program properties that were not discussed in [1], among which are the recognition network used, starting point of the gradient descent and different ways to partition style and content layers. We also give a brief comparison of some of the existing algorithm implementations and deep learning frameworks used. To study the style space further, an idea similar to [2] is used to generate synthetic images by maximizing a single entry in one of the Gram matrices G l and some interesting results are observed. Next, we try to mimic the sparsity and intensity distribution of Gram matrices obtained from a real painting and generate more complex textures. Finally, we propose two new style representations built on top of network's features and discuss how one could be used to achieve local and potentially content-aware style transfer.

arXiv (Cornell University), May 15, 2016

In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style... more In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style [7]. While showing great results when transferring homogeneous and repetitive patterns, the original style representation often fails to capture more complex properties, like having separate styles of foreground and background. This leads to visual artifacts and undesirable textures appearing in unexpected regions when performing style transfer. We tackle this issue with a variety of approaches, mostly by modifying the style representation in order for it to capture more information and impose a tighter constraint on the style transfer result. In our experiments, we subjectively evaluate our best method as producing from barely noticeable to significant improvements in the quality of style transfer.

Functional Imaging and Modelling of the Heart, 2017

As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in th... more As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in their translation to the bed side, in particular for abnormal cardiac rhythm diagnosis and therapy planning. However, the parameters that govern these models need to be estimated from noisy and sparse clinical data in an efficient and precise way, which is still an unsolved challenge. Invasive cardiac mapping provides the richest EP information available today. This paper proposes a new method to estimate a local map of electrical conductivities of the bi-ventricular heart by applying the back-propagation error concept, widely used in neural networks. The method works when either endocardial or epicardial activation time maps are available, and can cope with heterogeneous cardiac tissue. The method was evaluated on synthetic data, showing significantly increased performance in goodness of fit compared to a global parameter estimation approach. The resulting predictive power of the personalized model for cardiac resynchronization therapy was then assessed on 16 swine models of left bundle branch block with rich imaging and EP data before and after CRT. With the proposed personalization, the average error in activation time post CRT was \(10 \pm 4.5\) ms, lower than the observed pre/post-CRT difference of \(26.3 \pm 16.8\) ms.

JAMA Network Open, 2020

IMPORTANCE Mammography screening currently relies on subjective human interpretation. Artificial ... more IMPORTANCE Mammography screening currently relies on subjective human interpretation. Artificial intelligence (AI) advances could be used to increase mammography screening accuracy by reducing missed cancers and false positives. OBJECTIVE To evaluate whether AI can overcome human mammography interpretation limitations with a rigorous, unbiased evaluation of machine learning algorithms. DESIGN, SETTING, AND PARTICIPANTS In this diagnostic accuracy study conducted between September 2016 and November 2017, an international, crowdsourced challenge was hosted to foster AI algorithm development focused on interpreting screening mammography. More than 1100 participants comprising 126 teams from 44 countries participated. Analysis began November 18, 2016. MAIN OUTCOMES AND MEASUREMENTS Algorithms used images alone (challenge 1) or combined images, previous examinations (if available), and clinical and demographic risk factor data (challenge 2) and output a score that translated to cancer yes/no within 12 months. Algorithm accuracy for breast cancer detection was evaluated using area under the curve and algorithm specificity compared with radiologists' specificity with radiologists' sensitivity set at 85.9% (United States) and 83.9% (Sweden). An ensemble method aggregating top-performing AI algorithms and radiologists' recall assessment was developed and evaluated. RESULTS Overall, 144 231 screening mammograms from 85 580 US women (952 cancer positive Յ12 months from screening) were used for algorithm training and validation. A second independent validation cohort included 166 578 examinations from 68 008 Swedish women (780 cancer positive). The top-performing algorithm achieved an area under the curve of 0.858 (United States) and 0.903 (Sweden) and 66.2% (United States) and 81.2% (Sweden) specificity at the radiologists' sensitivity, lower than community-practice radiologists' specificity of 90.5% (United States) and 98.5% (Sweden). Combining top-performing algorithms and US radiologist assessments resulted in a higher area under the curve of 0.942 and achieved a significantly improved specificity (92.0%) at the same sensitivity. CONCLUSIONS AND RELEVANCE While no single AI algorithm outperformed radiologists, an ensemble of AI algorithms combined with radiologist assessment in a single-reader screening environment improved overall accuracy. This study underscores the potential of using machine (continued)

Medical Imaging 2017: Computer-Aided Diagnosis, 2017

A participant's experience of the Digital Mammography DREAM Challenge.

A participant's experience of the Digital Mammography DREAM Challenge.

arXiv (Cornell University), May 15, 2016

In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style... more In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style [7]. While showing great results when transferring homogeneous and repetitive patterns, the original style representation often fails to capture more complex properties, like having separate styles of foreground and background. This leads to visual artifacts and undesirable textures appearing in unexpected regions when performing style transfer. We tackle this issue with a variety of approaches, mostly by modifying the style representation in order for it to capture more information and impose a tighter constraint on the style transfer result. In our experiments, we subjectively evaluate our best method as producing from barely noticeable to significant improvements in the quality of style transfer.

Lecture Notes in Computer Science, 2017

As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in th... more As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in their translation to the bed side, in particular for abnormal cardiac rhythm diagnosis and therapy planning. However, the parameters that govern these models need to be estimated from noisy and sparse clinical data in an efficient and precise way, which is still an unsolved challenge. Invasive cardiac mapping provides the richest EP information available today. This paper proposes a new method to estimate a local map of electrical conductivities of the bi-ventricular heart by applying the back-propagation error concept, widely used in neural networks. The method works when either endocardial or epicardial activation time maps are available, and can cope with heterogeneous cardiac tissue. The method was evaluated on synthetic data, showing significantly increased performance in goodness of fit compared to a global parameter estimation approach. The resulting predictive power of the personalized model for cardiac resynchronization therapy was then assessed on 16 swine models of left bundle branch block with rich imaging and EP data before and after CRT. With the proposed personalization, the average error in activation time post CRT was \(10 \pm 4.5\) ms, lower than the observed pre/post-CRT difference of \(26.3 \pm 16.8\) ms.

arXiv (Cornell University), Feb 23, 2016

In this work we explore the method of style transfer presented in [1]. We first demonstrate the p... more In this work we explore the method of style transfer presented in [1]. We first demonstrate the power of the suggested style space on a few examples. We then vary different hyper-parameters and program properties that were not discussed in [1], among which are the recognition network used, starting point of the gradient descent and different ways to partition style and content layers. We also give a brief comparison of some of the existing algorithm implementations and deep learning frameworks used. To study the style space further, an idea similar to [2] is used to generate synthetic images by maximizing a single entry in one of the Gram matrices G l and some interesting results are observed. Next, we try to mimic the sparsity and intensity distribution of Gram matrices obtained from a real painting and generate more complex textures. Finally, we propose two new style representations built on top of network's features and discuss how one could be used to achieve local and potentially content-aware style transfer.

arXiv (Cornell University), Feb 23, 2016

In this work we explore the method of style transfer presented in [1]. We first demonstrate the p... more In this work we explore the method of style transfer presented in [1]. We first demonstrate the power of the suggested style space on a few examples. We then vary different hyper-parameters and program properties that were not discussed in [1], among which are the recognition network used, starting point of the gradient descent and different ways to partition style and content layers. We also give a brief comparison of some of the existing algorithm implementations and deep learning frameworks used. To study the style space further, an idea similar to [2] is used to generate synthetic images by maximizing a single entry in one of the Gram matrices G l and some interesting results are observed. Next, we try to mimic the sparsity and intensity distribution of Gram matrices obtained from a real painting and generate more complex textures. Finally, we propose two new style representations built on top of network's features and discuss how one could be used to achieve local and potentially content-aware style transfer.

arXiv (Cornell University), May 15, 2016

In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style... more In this work we investigate different avenues of improving the Neural Algorithm of Artistic Style [7]. While showing great results when transferring homogeneous and repetitive patterns, the original style representation often fails to capture more complex properties, like having separate styles of foreground and background. This leads to visual artifacts and undesirable textures appearing in unexpected regions when performing style transfer. We tackle this issue with a variety of approaches, mostly by modifying the style representation in order for it to capture more information and impose a tighter constraint on the style transfer result. In our experiments, we subjectively evaluate our best method as producing from barely noticeable to significant improvements in the quality of style transfer.

Functional Imaging and Modelling of the Heart, 2017

As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in th... more As models of cardiac electrophysiology (EP) are maturing, an increasing effort is being put in their translation to the bed side, in particular for abnormal cardiac rhythm diagnosis and therapy planning. However, the parameters that govern these models need to be estimated from noisy and sparse clinical data in an efficient and precise way, which is still an unsolved challenge. Invasive cardiac mapping provides the richest EP information available today. This paper proposes a new method to estimate a local map of electrical conductivities of the bi-ventricular heart by applying the back-propagation error concept, widely used in neural networks. The method works when either endocardial or epicardial activation time maps are available, and can cope with heterogeneous cardiac tissue. The method was evaluated on synthetic data, showing significantly increased performance in goodness of fit compared to a global parameter estimation approach. The resulting predictive power of the personalized model for cardiac resynchronization therapy was then assessed on 16 swine models of left bundle branch block with rich imaging and EP data before and after CRT. With the proposed personalization, the average error in activation time post CRT was \(10 \pm 4.5\) ms, lower than the observed pre/post-CRT difference of \(26.3 \pm 16.8\) ms.

JAMA Network Open, 2020

IMPORTANCE Mammography screening currently relies on subjective human interpretation. Artificial ... more IMPORTANCE Mammography screening currently relies on subjective human interpretation. Artificial intelligence (AI) advances could be used to increase mammography screening accuracy by reducing missed cancers and false positives. OBJECTIVE To evaluate whether AI can overcome human mammography interpretation limitations with a rigorous, unbiased evaluation of machine learning algorithms. DESIGN, SETTING, AND PARTICIPANTS In this diagnostic accuracy study conducted between September 2016 and November 2017, an international, crowdsourced challenge was hosted to foster AI algorithm development focused on interpreting screening mammography. More than 1100 participants comprising 126 teams from 44 countries participated. Analysis began November 18, 2016. MAIN OUTCOMES AND MEASUREMENTS Algorithms used images alone (challenge 1) or combined images, previous examinations (if available), and clinical and demographic risk factor data (challenge 2) and output a score that translated to cancer yes/no within 12 months. Algorithm accuracy for breast cancer detection was evaluated using area under the curve and algorithm specificity compared with radiologists' specificity with radiologists' sensitivity set at 85.9% (United States) and 83.9% (Sweden). An ensemble method aggregating top-performing AI algorithms and radiologists' recall assessment was developed and evaluated. RESULTS Overall, 144 231 screening mammograms from 85 580 US women (952 cancer positive Յ12 months from screening) were used for algorithm training and validation. A second independent validation cohort included 166 578 examinations from 68 008 Swedish women (780 cancer positive). The top-performing algorithm achieved an area under the curve of 0.858 (United States) and 0.903 (Sweden) and 66.2% (United States) and 81.2% (Sweden) specificity at the radiologists' sensitivity, lower than community-practice radiologists' specificity of 90.5% (United States) and 98.5% (Sweden). Combining top-performing algorithms and US radiologist assessments resulted in a higher area under the curve of 0.942 and achieved a significantly improved specificity (92.0%) at the same sensitivity. CONCLUSIONS AND RELEVANCE While no single AI algorithm outperformed radiologists, an ensemble of AI algorithms combined with radiologist assessment in a single-reader screening environment improved overall accuracy. This study underscores the potential of using machine (continued)

Medical Imaging 2017: Computer-Aided Diagnosis, 2017

A participant's experience of the Digital Mammography DREAM Challenge.