Two-Stage Segmentation Framework Based on Distance Transformation
With the rise of deep learning, using deep learning to segment lesions and assist in diagnosis has become an effective means to promote clinical medical analysis. However, the partial volume effect of organ tissues leads to unclear and blurred edges of ROI in medical images, making it challenging to...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-12-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/22/1/250 |
| _version_ | 1797497548287508480 |
|---|---|
| author | Xiaoyang Huang Zhi Lin Yudi Jiao Moon-Tong Chan Shaohui Huang Liansheng Wang |
| author_facet | Xiaoyang Huang Zhi Lin Yudi Jiao Moon-Tong Chan Shaohui Huang Liansheng Wang |
| author_sort | Xiaoyang Huang |
| collection | DOAJ |
| description | With the rise of deep learning, using deep learning to segment lesions and assist in diagnosis has become an effective means to promote clinical medical analysis. However, the partial volume effect of organ tissues leads to unclear and blurred edges of ROI in medical images, making it challenging to achieve high-accuracy segmentation of lesions or organs. In this paper, we assume that the distance map obtained by performing distance transformation on the ROI edge can be used as a weight map to make the network pay more attention to the learning of the ROI edge region. To this end, we design a novel framework to flexibly embed the distance map into the two-stage network to improve left atrium MRI segmentation performance. Furthermore, a series of distance map generation methods are proposed and studied to reasonably explore how to express the weight of assisting network learning. We conduct thorough experiments to verify the effectiveness of the proposed segmentation framework, and experimental results demonstrate that our hypothesis is feasible. |
| first_indexed | 2024-03-10T03:20:47Z |
| format | Article |
| id | doaj.art-160594b5331e4ca081fdfe0d86530a91 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T03:20:47Z |
| publishDate | 2021-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-160594b5331e4ca081fdfe0d86530a912023-11-23T12:19:21ZengMDPI AGSensors1424-82202021-12-0122125010.3390/s22010250Two-Stage Segmentation Framework Based on Distance TransformationXiaoyang Huang0Zhi Lin1Yudi Jiao2Moon-Tong Chan3Shaohui Huang4Liansheng Wang5Department of Computer Science, School of Informatics, Xiamen University, Xiamen 361005, ChinaDepartment of Computer Science, School of Informatics, Xiamen University, Xiamen 361005, ChinaDepartment of Computer Science, School of Informatics, Xiamen University, Xiamen 361005, ChinaSchool of Science and Technology, Hong Kong Metropolitan University, Homantin, Kowloon 999077, Hong KongDepartment of Computer Science, School of Informatics, Xiamen University, Xiamen 361005, ChinaDepartment of Computer Science, School of Informatics, Xiamen University, Xiamen 361005, ChinaWith the rise of deep learning, using deep learning to segment lesions and assist in diagnosis has become an effective means to promote clinical medical analysis. However, the partial volume effect of organ tissues leads to unclear and blurred edges of ROI in medical images, making it challenging to achieve high-accuracy segmentation of lesions or organs. In this paper, we assume that the distance map obtained by performing distance transformation on the ROI edge can be used as a weight map to make the network pay more attention to the learning of the ROI edge region. To this end, we design a novel framework to flexibly embed the distance map into the two-stage network to improve left atrium MRI segmentation performance. Furthermore, a series of distance map generation methods are proposed and studied to reasonably explore how to express the weight of assisting network learning. We conduct thorough experiments to verify the effectiveness of the proposed segmentation framework, and experimental results demonstrate that our hypothesis is feasible.https://www.mdpi.com/1424-8220/22/1/250distance transformationtwo-stagedeep learningmedical image segmentation |
| spellingShingle | Xiaoyang Huang Zhi Lin Yudi Jiao Moon-Tong Chan Shaohui Huang Liansheng Wang Two-Stage Segmentation Framework Based on Distance Transformation Sensors distance transformation two-stage deep learning medical image segmentation |
| title | Two-Stage Segmentation Framework Based on Distance Transformation |
| title_full | Two-Stage Segmentation Framework Based on Distance Transformation |
| title_fullStr | Two-Stage Segmentation Framework Based on Distance Transformation |
| title_full_unstemmed | Two-Stage Segmentation Framework Based on Distance Transformation |
| title_short | Two-Stage Segmentation Framework Based on Distance Transformation |
| title_sort | two stage segmentation framework based on distance transformation |
| topic | distance transformation two-stage deep learning medical image segmentation |
| url | https://www.mdpi.com/1424-8220/22/1/250 |
| work_keys_str_mv | AT xiaoyanghuang twostagesegmentationframeworkbasedondistancetransformation AT zhilin twostagesegmentationframeworkbasedondistancetransformation AT yudijiao twostagesegmentationframeworkbasedondistancetransformation AT moontongchan twostagesegmentationframeworkbasedondistancetransformation AT shaohuihuang twostagesegmentationframeworkbasedondistancetransformation AT lianshengwang twostagesegmentationframeworkbasedondistancetransformation |