Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9433861
Ivana Kojcic, T. Papadopoulo, R. Deriche, Samuel Deslauriers-Gauthier
White matter fibers transfer the information between brain regions with delays that are measurable with magnetoencephalography and electroencephalography (M/EEG). In the context of regularizing the dynamics of M/EEG and recovering electrical activity of the brain from M/EEG measurements, this article proposes a graph representation-based framework to solve the M/EEG inverse problem, where prior information about transmission delays supported by diffusion MRI (dMRI) are included to enforce temporal smoothness. Results of the reconstruction of brain activity from simulated MEG measurements are compared to MNE, LORETA and CGS methods and we show that our approach improves MEG source localization when compared to these three state-of-the-art approaches. In addition, we show preliminary qualitative results of the proposed reconstruction method on real MEG data for a sensory-motor task.
{"title":"Incorporating Transmission Delays Supported By Diffusion Mri In Meg Source Reconstruction","authors":"Ivana Kojcic, T. Papadopoulo, R. Deriche, Samuel Deslauriers-Gauthier","doi":"10.1109/ISBI48211.2021.9433861","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9433861","url":null,"abstract":"White matter fibers transfer the information between brain regions with delays that are measurable with magnetoencephalography and electroencephalography (M/EEG). In the context of regularizing the dynamics of M/EEG and recovering electrical activity of the brain from M/EEG measurements, this article proposes a graph representation-based framework to solve the M/EEG inverse problem, where prior information about transmission delays supported by diffusion MRI (dMRI) are included to enforce temporal smoothness. Results of the reconstruction of brain activity from simulated MEG measurements are compared to MNE, LORETA and CGS methods and we show that our approach improves MEG source localization when compared to these three state-of-the-art approaches. In addition, we show preliminary qualitative results of the proposed reconstruction method on real MEG data for a sensory-motor task.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125018199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9433902
Yuhui Du, Ju Niu, V. Calhoun
It is difficult to distinguish schizophrenia (SZ) and bipolar disorder with psychosis (BPP) due to their overlapping symptoms. Indeed, there has been evidence supporting different subtypes within them. Data-driven clustering approaches are commonly used to explore biologically meaningful biotypes using neuroimaging features. However, previous studies typically consider pair-wise subject relationships. Here, we propose a hypergraph clustering method to explore biotypes. Our method extracts high-order features via hyperedges sampling, measures similarity and then regroups subjects using community detection. We applied it to identify biotypes of 100 BPP and 100 SZ patients using brain functional connectivity estimated from resting-state fMRI data, and compared with solutions from K-means and normalized cut (Ncut). Two reliable biotypes were identified and had greater differences in functional connectivity than groups determined by clinical diagnosis. Our method also outperformed K-means and Ncut for the clustering ability and computation efficiency. In summary, the proposed method is promising for developing biotypes, targeting accurate clinical diagnosis for psychosis.
{"title":"A New Hypergraph Clustering Method For Exploring Transdiagnostic Biotypes In Mental Illnesses: Application To Schizophrenia And Psychotic Bipolar Disorder","authors":"Yuhui Du, Ju Niu, V. Calhoun","doi":"10.1109/ISBI48211.2021.9433902","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9433902","url":null,"abstract":"It is difficult to distinguish schizophrenia (SZ) and bipolar disorder with psychosis (BPP) due to their overlapping symptoms. Indeed, there has been evidence supporting different subtypes within them. Data-driven clustering approaches are commonly used to explore biologically meaningful biotypes using neuroimaging features. However, previous studies typically consider pair-wise subject relationships. Here, we propose a hypergraph clustering method to explore biotypes. Our method extracts high-order features via hyperedges sampling, measures similarity and then regroups subjects using community detection. We applied it to identify biotypes of 100 BPP and 100 SZ patients using brain functional connectivity estimated from resting-state fMRI data, and compared with solutions from K-means and normalized cut (Ncut). Two reliable biotypes were identified and had greater differences in functional connectivity than groups determined by clinical diagnosis. Our method also outperformed K-means and Ncut for the clustering ability and computation efficiency. In summary, the proposed method is promising for developing biotypes, targeting accurate clinical diagnosis for psychosis.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129352044","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9434111
Kaveri A. Thakoor, Darius D Bordbar, Jiaang Yao, Omar Moussa, R. Chen, P. Sajda
With the availability and increasing reliance on the noninvasive Optical Coherence Tomography Angiography(OCTA) imaging technique for detection of vascular diseases of the retina,suchasage-related macular degeneration(AMD),clinicians now have access to more data than they can effectively parse and digest. Artificial intelligence in the form of convolutional neural networks (CNNs), have shown successful detection of AMDvs. no AMD from fundus images as well as from OCT structural images. In this work, we address an ovel classification problem: automated detection of late stage of the disease, neovascular AMD, visualized through presence of choroidal neovascularization (CNV) and its sequelae. We describe hybrid 3D-2D CNNs that achieve accuracy up to 77.8% at multi-class categorical classification of non-AMD eyes, eyes having non-neovascular AMD, and eyes having neovascular AMD, offering a first-of-its-kind deep learning approach for differentiating progression in AMD.
{"title":"Hybrid 3d-2d Deep Learning For Detection Of Neovascularage-Related Macular Degeneration Using Optical Coherence Tomography B-Scans And Angiography Volumes","authors":"Kaveri A. Thakoor, Darius D Bordbar, Jiaang Yao, Omar Moussa, R. Chen, P. Sajda","doi":"10.1109/ISBI48211.2021.9434111","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9434111","url":null,"abstract":"With the availability and increasing reliance on the noninvasive Optical Coherence Tomography Angiography(OCTA) imaging technique for detection of vascular diseases of the retina,suchasage-related macular degeneration(AMD),clinicians now have access to more data than they can effectively parse and digest. Artificial intelligence in the form of convolutional neural networks (CNNs), have shown successful detection of AMDvs. no AMD from fundus images as well as from OCT structural images. In this work, we address an ovel classification problem: automated detection of late stage of the disease, neovascular AMD, visualized through presence of choroidal neovascularization (CNV) and its sequelae. We describe hybrid 3D-2D CNNs that achieve accuracy up to 77.8% at multi-class categorical classification of non-AMD eyes, eyes having non-neovascular AMD, and eyes having neovascular AMD, offering a first-of-its-kind deep learning approach for differentiating progression in AMD.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"143 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129630478","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9434144
Sumit Kaushik, J. Kybic, Avinash Bansal, Temesgen Bihonegn, J. Slovák
In this paper, we provide a framework to evaluate new scalar quantities for higher order tensors (HOT) appearing in high angular resolution diffusion imaging (HARDI). These can potentially serve as biomarkers. It involves flattening of HOTs and extraction of the diagonal D-components. Experiments performed in the 4th order case reveal that D-components encode geometric information unlike the isometric 6D 2nd order Voigt form. The existing invariants obtained from the Voigt form are considered for comparison. We also notice that D-components can be useful in segmentation of white matter structures in crossing regions and classification. Results on phantom and the synthetic dataset support the conclusions.
{"title":"Potential Biomarkers From Positive Definite 4th Order Tensors In Hardi","authors":"Sumit Kaushik, J. Kybic, Avinash Bansal, Temesgen Bihonegn, J. Slovák","doi":"10.1109/ISBI48211.2021.9434144","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9434144","url":null,"abstract":"In this paper, we provide a framework to evaluate new scalar quantities for higher order tensors (HOT) appearing in high angular resolution diffusion imaging (HARDI). These can potentially serve as biomarkers. It involves flattening of HOTs and extraction of the diagonal D-components. Experiments performed in the 4th order case reveal that D-components encode geometric information unlike the isometric 6D 2nd order Voigt form. The existing invariants obtained from the Voigt form are considered for comparison. We also notice that D-components can be useful in segmentation of white matter structures in crossing regions and classification. Results on phantom and the synthetic dataset support the conclusions.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122947677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9434072
Mourad Gridach, I. Voiculescu
Since manual annotation of medical images is time consuming for clinical experts, reliable automatic segmentation would be the ideal way to handle large medical datasets. Deep learning-based models have been the dominant approach, achieving remarkable performance on various medical segmentation tasks. There can be a significant variation in the size of the feature being segmented out of a medical image relative to the other features in the image, which can be challenging. In this paper, we propose a Densely Oriented Pooling Network (DOPNet) to capture variation in feature size in medical images and preserve spatial interconnection. DOPNet is based on two interdependent ideas: the dense connectivity and the pooling oriented layer. When tested on three publicly available medical image segmentation datasets, the proposed model achieves leading performance.
{"title":"Dopnet: Densely Oriented Pooling Network For Medical Image Segmentation","authors":"Mourad Gridach, I. Voiculescu","doi":"10.1109/ISBI48211.2021.9434072","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9434072","url":null,"abstract":"Since manual annotation of medical images is time consuming for clinical experts, reliable automatic segmentation would be the ideal way to handle large medical datasets. Deep learning-based models have been the dominant approach, achieving remarkable performance on various medical segmentation tasks. There can be a significant variation in the size of the feature being segmented out of a medical image relative to the other features in the image, which can be challenging. In this paper, we propose a Densely Oriented Pooling Network (DOPNet) to capture variation in feature size in medical images and preserve spatial interconnection. DOPNet is based on two interdependent ideas: the dense connectivity and the pooling oriented layer. When tested on three publicly available medical image segmentation datasets, the proposed model achieves leading performance.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122976033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9433759
Christian Ritter, Roman Spilger, Ji Young Lee, R. Bartenschlager, K. Rohr
Tracking of subcellular structures displayed as spots in fluorescence microscopy images is important to quantify viral and cellular processes. We have developed a novel tracking approach for biological particles which uses deep learning for both particle detection and particle association. Our approach combines a domain adapted Deconvolution Network for particle detection with an LSTM-based recurrent neural network for tracking. Past and future information in both forward and backward direction is exploited by bidirectional LSTMs, and assignment probabilities are determined jointly across multiple detections. We evaluated the proposed approach using image sequences of the Particle Tracking Challenge as well as live cell fluorescence microscopy data of hepatitis C virus proteins. It turned out that our approach yields state-of-the-art results or improves the results compared to previous methods.
{"title":"Deep Learning For Particle Detection And Tracking In Fluorescence Microscopy Images","authors":"Christian Ritter, Roman Spilger, Ji Young Lee, R. Bartenschlager, K. Rohr","doi":"10.1109/ISBI48211.2021.9433759","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9433759","url":null,"abstract":"Tracking of subcellular structures displayed as spots in fluorescence microscopy images is important to quantify viral and cellular processes. We have developed a novel tracking approach for biological particles which uses deep learning for both particle detection and particle association. Our approach combines a domain adapted Deconvolution Network for particle detection with an LSTM-based recurrent neural network for tracking. Past and future information in both forward and backward direction is exploited by bidirectional LSTMs, and assignment probabilities are determined jointly across multiple detections. We evaluated the proposed approach using image sequences of the Particle Tracking Challenge as well as live cell fluorescence microscopy data of hepatitis C virus proteins. It turned out that our approach yields state-of-the-art results or improves the results compared to previous methods.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"434 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126098839","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9434040
M. Beetz, Abhirup Banerjee, V. Grau
Many important cardiac biomarkers used in clinical practice describe cardiac anatomy and function in three dimensions (3D). However, common cardiac magnetic resonance imaging (MRI) protocols often only generate two-dimensional (2D) image slices of the underlying 3D anatomy and are susceptible to various types of motion artifacts causing slice misalignment. In this paper, we propose a deep learning method acting directly on point clouds to reconstruct a dense 3D biventricular heart model from misaligned 2D cardiac MR image contours. The method is able to reduce mild, medium, and strong slice misalignments (mean translation $sim 3.5$ mm; mean rotation $sim 2.5^{circ})$ to a Chamfer distance below image resolution (1.25 mm) with high robustness (standard deviation 0.18 mm) on a statistical shape model dataset. It also manages to reconstruct smooth 3D shapes with accurate left ventricular volumes from cine MR images of the UK Biobank study.
{"title":"Biventricular Surface Reconstruction From Cine Mri Contours Using Point Completion Networks","authors":"M. Beetz, Abhirup Banerjee, V. Grau","doi":"10.1109/ISBI48211.2021.9434040","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9434040","url":null,"abstract":"Many important cardiac biomarkers used in clinical practice describe cardiac anatomy and function in three dimensions (3D). However, common cardiac magnetic resonance imaging (MRI) protocols often only generate two-dimensional (2D) image slices of the underlying 3D anatomy and are susceptible to various types of motion artifacts causing slice misalignment. In this paper, we propose a deep learning method acting directly on point clouds to reconstruct a dense 3D biventricular heart model from misaligned 2D cardiac MR image contours. The method is able to reduce mild, medium, and strong slice misalignments (mean translation $sim 3.5$ mm; mean rotation $sim 2.5^{circ})$ to a Chamfer distance below image resolution (1.25 mm) with high robustness (standard deviation 0.18 mm) on a statistical shape model dataset. It also manages to reconstruct smooth 3D shapes with accurate left ventricular volumes from cine MR images of the UK Biobank study.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114173448","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9433841
A. Naglah, F. Khalifa, R. Khaled, A. Razek, A. El-Baz
Achieving early detection and classification of thyroid nodules contributes to the prediction of cancer burdening and also steers appropriate clinical pathways of that medical condition. We propose a novel multimodal MRI-based computer-aided diagnosis (CAD) system that detects cancerous thyroid nodules using a deep-learning architecture. Particularly, our system is built with a multi-input convolutional neural network (CNN) to perform fusion of two MRI modalities: the diffusion weighted image (DWI) and apparent diffusion coefficient (ADC) map. The main contribution of our system is three-folded. Namely, (1) it is the first system to fuse thyroid DWI and ADC using CNN for classification purpose; (2) it enables independent convolutions process for each of DWI and ADC images, which can increase the likelihood of detecting deep texture patterns in thyroid nodules; and (3) it enables adding extra channels in each input with the possibility to integrate with additional MRI modalities and other imaging technologies. We compared our system to other fusion methods and also to other machine learning (ML) frameworks that use hand-crafted features. Our system achieved the highest performance among them with diagnostic accuracy of 0.88, precision of 0.82, and recall of 0.82.
{"title":"Thyroid Cancer Computer-Aided Diagnosis System using MRI-Based Multi-Input CNN Model","authors":"A. Naglah, F. Khalifa, R. Khaled, A. Razek, A. El-Baz","doi":"10.1109/ISBI48211.2021.9433841","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9433841","url":null,"abstract":"Achieving early detection and classification of thyroid nodules contributes to the prediction of cancer burdening and also steers appropriate clinical pathways of that medical condition. We propose a novel multimodal MRI-based computer-aided diagnosis (CAD) system that detects cancerous thyroid nodules using a deep-learning architecture. Particularly, our system is built with a multi-input convolutional neural network (CNN) to perform fusion of two MRI modalities: the diffusion weighted image (DWI) and apparent diffusion coefficient (ADC) map. The main contribution of our system is three-folded. Namely, (1) it is the first system to fuse thyroid DWI and ADC using CNN for classification purpose; (2) it enables independent convolutions process for each of DWI and ADC images, which can increase the likelihood of detecting deep texture patterns in thyroid nodules; and (3) it enables adding extra channels in each input with the possibility to integrate with additional MRI modalities and other imaging technologies. We compared our system to other fusion methods and also to other machine learning (ML) frameworks that use hand-crafted features. Our system achieved the highest performance among them with diagnostic accuracy of 0.88, precision of 0.82, and recall of 0.82.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125684654","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9433850
Qin Wang, Weibing Zhao, Xu Yan, Hui Che, Kunlin Ye, Yingfeng Lu, Zhen Li, Shuguang Cui
Accurate semantic segmentation of coronary artery for CT images is critical in both coronary-related disease diagnosis (e.g., stenosis detection and plaque grading) and further intervention treatments. Considering the irrelevant tubular structures are usually difficult to be distinguished from the coronary arteries, e.g., veins, existing methods inevitably lead to false positives. In this paper, we incorporate the voxel and point cloud based segmentation methods into a coarse-to-fine framework for accurate coronary artery segmentation from Coronary Computed Tomography Angiography (CCTA) images. Specifically, after the coarse segmentation from any appealing voxel-based framework, initial segmentation maps are converted into point clouds and fed into a Refinement Module to filter out the irrelevant tubular vessels. In practice, the Refinement Module adopts the local feature aggregation on point clouds for contextual learning, capturing the geometric morphology of the coronary arteries. Furthermore, the first annotated CCTA dataset for coronary artery segmentation, named CORONARY-481, is released in this paper. Extensive experiments indicate that the proposed approach achieves state-of-the-art performance in coronary artery segmentation, improving the dice metric by 10% and preserving its fine structure as well.
{"title":"Geometric Morphology Based Irrelevant Vessels Removal For Accurate Coronary Artery Segmentation","authors":"Qin Wang, Weibing Zhao, Xu Yan, Hui Che, Kunlin Ye, Yingfeng Lu, Zhen Li, Shuguang Cui","doi":"10.1109/ISBI48211.2021.9433850","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9433850","url":null,"abstract":"Accurate semantic segmentation of coronary artery for CT images is critical in both coronary-related disease diagnosis (e.g., stenosis detection and plaque grading) and further intervention treatments. Considering the irrelevant tubular structures are usually difficult to be distinguished from the coronary arteries, e.g., veins, existing methods inevitably lead to false positives. In this paper, we incorporate the voxel and point cloud based segmentation methods into a coarse-to-fine framework for accurate coronary artery segmentation from Coronary Computed Tomography Angiography (CCTA) images. Specifically, after the coarse segmentation from any appealing voxel-based framework, initial segmentation maps are converted into point clouds and fed into a Refinement Module to filter out the irrelevant tubular vessels. In practice, the Refinement Module adopts the local feature aggregation on point clouds for contextual learning, capturing the geometric morphology of the coronary arteries. Furthermore, the first annotated CCTA dataset for coronary artery segmentation, named CORONARY-481, is released in this paper. Extensive experiments indicate that the proposed approach achieves state-of-the-art performance in coronary artery segmentation, improving the dice metric by 10% and preserving its fine structure as well.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131403647","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-04-13DOI: 10.1109/ISBI48211.2021.9433859
Chunyu Dong, Qunfei Zhao, Kun Chen, Xiaolin Huang
Institute of Medical Robotics, Shanghai Jiao Tong University, Shanghai, China Rescaling a feature map could be a key issue of biological image segmentation. Nearly all the existing upsampling strategies concentrate only on local information and attempt to expand the reception field via deepening the network. In this paper, we present Asymmetric Attention Upsampling (AAU) for biological-image segmentation. AAU utilizes the information of low-level feature maps to rescale the high-level feature maps smartly through spatial pooling and attention mechanisms. It consists of two attention variants: Asymmetric Spatial Attention (ASA) and Asymmetric Channel Attention (ACA). The Asymmetric Attention Upsampling Network (AAU-Net) combines several AAU blocks to achieve better segmentation performance. Experiments on the Kvasir-SEG data set reveal the effectiveness of our work. AAU-Net outperforms other state-of-the-art methods for polyp segmentation while not consuming many resources.
{"title":"Asymmetric Attention Upsampling: Rethinking Upsampling For Biological Image Segmentation","authors":"Chunyu Dong, Qunfei Zhao, Kun Chen, Xiaolin Huang","doi":"10.1109/ISBI48211.2021.9433859","DOIUrl":"https://doi.org/10.1109/ISBI48211.2021.9433859","url":null,"abstract":"Institute of Medical Robotics, Shanghai Jiao Tong University, Shanghai, China Rescaling a feature map could be a key issue of biological image segmentation. Nearly all the existing upsampling strategies concentrate only on local information and attempt to expand the reception field via deepening the network. In this paper, we present Asymmetric Attention Upsampling (AAU) for biological-image segmentation. AAU utilizes the information of low-level feature maps to rescale the high-level feature maps smartly through spatial pooling and attention mechanisms. It consists of two attention variants: Asymmetric Spatial Attention (ASA) and Asymmetric Channel Attention (ACA). The Asymmetric Attention Upsampling Network (AAU-Net) combines several AAU blocks to achieve better segmentation performance. Experiments on the Kvasir-SEG data set reveal the effectiveness of our work. AAU-Net outperforms other state-of-the-art methods for polyp segmentation while not consuming many resources.","PeriodicalId":372939,"journal":{"name":"2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131541947","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：