Improving Whole-Heart CT Image Segmentation by Attention Mechanism
Decent whole-heart segmentation from computed tomography (CT) can greatly contribute to the diagnosis and treatment of cardiovascular diseases. However, due to the difficulties such as blurred boundaries between neighbouring tissues and a large number of background voxels in medical images, automate...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8938714/ |
| _version_ | 1818853284836802560 |
|---|---|
| author | Wei Wang Chengqin Ye Shanzhuo Zhang Yong Xu Kuanquan Wang |
| author_facet | Wei Wang Chengqin Ye Shanzhuo Zhang Yong Xu Kuanquan Wang |
| author_sort | Wei Wang |
| collection | DOAJ |
| description | Decent whole-heart segmentation from computed tomography (CT) can greatly contribute to the diagnosis and treatment of cardiovascular diseases. However, due to the difficulties such as blurred boundaries between neighbouring tissues and a large number of background voxels in medical images, automated whole-heart segmentation is still a challenging task. In this paper, we proposed three modified attention models, including simple negative example mining (SNEM), attention gate (AG) and U-CliqueNet (UCNet), to lead the deep learning network to focus on more salient information. These three attention modules were further implemented into a deeply-supervised 3D UNET separately and jointly, showing different degrees of improvement on the whole-heart segmentation task. Our experiments advised that SNEM was the most simple and effective attention mechanism for medical image processing among the three and the UCNet could reach the best performance. The combination of the attention mechanisms cannot always synergistically increase the accuracy, but joint models would have a positive influence in most cases. Finally, our network achieved a Dice score of 0.9112, which was a substantially higher performance than most of the state-of-the-art methods. |
| first_indexed | 2024-12-19T07:34:23Z |
| format | Article |
| id | doaj.art-35f53633dffe4cf7baf227c9c43080f6 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-19T07:34:23Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-35f53633dffe4cf7baf227c9c43080f62022-12-21T20:30:37ZengIEEEIEEE Access2169-35362020-01-018145791458710.1109/ACCESS.2019.29614108938714Improving Whole-Heart CT Image Segmentation by Attention MechanismWei Wang0https://orcid.org/0000-0002-1874-9947Chengqin Ye1https://orcid.org/0000-0002-3228-0524Shanzhuo Zhang2https://orcid.org/0000-0002-0098-8536Yong Xu3https://orcid.org/0000-0003-0530-2123Kuanquan Wang4https://orcid.org/0000-0003-1347-3491Bio-Computing Research Center, Harbin Institute of Technology, Shenzhen, ChinaSchool of Computer Science and Technology, Harbin Institute of Technology, Harbin, ChinaSchool of Computer Science and Technology, Harbin Institute of Technology, Harbin, ChinaBio-Computing Research Center, Harbin Institute of Technology, Shenzhen, ChinaSchool of Computer Science and Technology, Harbin Institute of Technology, Harbin, ChinaDecent whole-heart segmentation from computed tomography (CT) can greatly contribute to the diagnosis and treatment of cardiovascular diseases. However, due to the difficulties such as blurred boundaries between neighbouring tissues and a large number of background voxels in medical images, automated whole-heart segmentation is still a challenging task. In this paper, we proposed three modified attention models, including simple negative example mining (SNEM), attention gate (AG) and U-CliqueNet (UCNet), to lead the deep learning network to focus on more salient information. These three attention modules were further implemented into a deeply-supervised 3D UNET separately and jointly, showing different degrees of improvement on the whole-heart segmentation task. Our experiments advised that SNEM was the most simple and effective attention mechanism for medical image processing among the three and the UCNet could reach the best performance. The combination of the attention mechanisms cannot always synergistically increase the accuracy, but joint models would have a positive influence in most cases. Finally, our network achieved a Dice score of 0.9112, which was a substantially higher performance than most of the state-of-the-art methods.https://ieeexplore.ieee.org/document/8938714/Medical image processingCT image segmentationattention mechanismattention gatefeedback connection |
| spellingShingle | Wei Wang Chengqin Ye Shanzhuo Zhang Yong Xu Kuanquan Wang Improving Whole-Heart CT Image Segmentation by Attention Mechanism IEEE Access Medical image processing CT image segmentation attention mechanism attention gate feedback connection |
| title | Improving Whole-Heart CT Image Segmentation by Attention Mechanism |
| title_full | Improving Whole-Heart CT Image Segmentation by Attention Mechanism |
| title_fullStr | Improving Whole-Heart CT Image Segmentation by Attention Mechanism |
| title_full_unstemmed | Improving Whole-Heart CT Image Segmentation by Attention Mechanism |
| title_short | Improving Whole-Heart CT Image Segmentation by Attention Mechanism |
| title_sort | improving whole heart ct image segmentation by attention mechanism |
| topic | Medical image processing CT image segmentation attention mechanism attention gate feedback connection |
| url | https://ieeexplore.ieee.org/document/8938714/ |
| work_keys_str_mv | AT weiwang improvingwholeheartctimagesegmentationbyattentionmechanism AT chengqinye improvingwholeheartctimagesegmentationbyattentionmechanism AT shanzhuozhang improvingwholeheartctimagesegmentationbyattentionmechanism AT yongxu improvingwholeheartctimagesegmentationbyattentionmechanism AT kuanquanwang improvingwholeheartctimagesegmentationbyattentionmechanism |