MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural image
Scoliosis is among the most prevalent diseases affecting teenagers. However, traditional scoliosis screening methods often resort to physical examination or radiographic imaging. The two ways both rely on experts with high costs, which are not suitable for wide-range screening. Besides, estimating C...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-01-01
|
| Series: | Alexandria Engineering Journal |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1110016824012225 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832595910258851840 |
|---|---|
| author | Xiaojia Zhu Rui Chen Zhiwen Shao Ming Zhang Yuhu Dai Wenzhi Zhang Chuandong Lang |
| author_facet | Xiaojia Zhu Rui Chen Zhiwen Shao Ming Zhang Yuhu Dai Wenzhi Zhang Chuandong Lang |
| author_sort | Xiaojia Zhu |
| collection | DOAJ |
| description | Scoliosis is among the most prevalent diseases affecting teenagers. However, traditional scoliosis screening methods often resort to physical examination or radiographic imaging. The two ways both rely on experts with high costs, which are not suitable for wide-range screening. Besides, estimating Cobb angle level only using natural images are challenging. To tackle these issues, we propose a multi-grained scoliosis detection framework by jointly estimating severity level and Cobb angle level of scoliosis from a natural image instead of a radiographic image, which has not been explored before. Specifically, we regard scoliosis estimation as an ordinal regression problem, and transform it into a series of binary classification sub-problems. Besides, we adopt the visual attention network with large kernel attention as the backbone for feature learning, which can model local and global correlations with efficient computations. The feature learning and the ordinal regression is put into an end-to-end framework, in which the two tasks of scoliosis severity level estimation and scoliosis angle level estimation are jointly learned and can contribute to each other. Extensive experiments demonstrate that our approach outperforms state-of-the-art methods as well as human performance, which provides a promising and economical solution to wide-range scoliosis screening. Particularly, our approach achieves accuracies of 94.90% and 79.62% in estimating severity level and Cobb angle level, improving large margins of 4.90% and 12.15% over existing natural image based scoliosis detection performance, respectively. The code is available at https://github.com/RuiChen-stack/MGScoliosis. |
| format | Article |
| id | doaj-art-2aee0251bbac4174809bb19640c53826 |
| institution | Kabale University |
| issn | 1110-0168 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Alexandria Engineering Journal |
| spelling | doaj-art-2aee0251bbac4174809bb19640c538262025-01-18T05:03:40ZengElsevierAlexandria Engineering Journal1110-01682025-01-01111329340MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural imageXiaojia Zhu0Rui Chen1Zhiwen Shao2Ming Zhang3Yuhu Dai4Wenzhi Zhang5Chuandong Lang6Department of Orthopedics, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei 230001, China; Xuzhou Central Hospital/The Xuzhou Clinical College of Xuzhou Medical University, Xuzhou 221009, China; Xuzhou Rehabilitation Hospital/The Affiliated Xuzhou Rehabilitation Hospital of Xuzhou Medical University, Xuzhou 221003, ChinaSchool of Computer Science and Technology, China University of Mining and Technology, Xuzhou 221116, ChinaDepartment of Orthopedics, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei 230001, China; School of Computer Science and Technology, China University of Mining and Technology, Xuzhou 221116, ChinaXuzhou Central Hospital/The Xuzhou Clinical College of Xuzhou Medical University, Xuzhou 221009, China; Xuzhou Rehabilitation Hospital/The Affiliated Xuzhou Rehabilitation Hospital of Xuzhou Medical University, Xuzhou 221003, ChinaDepartment of Orthopedics, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei 230001, China; Department of Orthopaedic Surgery, The First Affiliated Hospital, Sun Yat-Sen University, Guangzhou 510080, ChinaDepartment of Orthopedics, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei 230001, ChinaDepartment of Orthopedics, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei 230001, China; Department of Orthopaedic Surgery, The First Affiliated Hospital, Sun Yat-Sen University, Guangzhou 510080, China; Corresponding author at: Department of Orthopedics, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei 230001, China.Scoliosis is among the most prevalent diseases affecting teenagers. However, traditional scoliosis screening methods often resort to physical examination or radiographic imaging. The two ways both rely on experts with high costs, which are not suitable for wide-range screening. Besides, estimating Cobb angle level only using natural images are challenging. To tackle these issues, we propose a multi-grained scoliosis detection framework by jointly estimating severity level and Cobb angle level of scoliosis from a natural image instead of a radiographic image, which has not been explored before. Specifically, we regard scoliosis estimation as an ordinal regression problem, and transform it into a series of binary classification sub-problems. Besides, we adopt the visual attention network with large kernel attention as the backbone for feature learning, which can model local and global correlations with efficient computations. The feature learning and the ordinal regression is put into an end-to-end framework, in which the two tasks of scoliosis severity level estimation and scoliosis angle level estimation are jointly learned and can contribute to each other. Extensive experiments demonstrate that our approach outperforms state-of-the-art methods as well as human performance, which provides a promising and economical solution to wide-range scoliosis screening. Particularly, our approach achieves accuracies of 94.90% and 79.62% in estimating severity level and Cobb angle level, improving large margins of 4.90% and 12.15% over existing natural image based scoliosis detection performance, respectively. The code is available at https://github.com/RuiChen-stack/MGScoliosis.http://www.sciencedirect.com/science/article/pii/S1110016824012225Computer visionArtificial intelligenceScoliosis detectionJoint learningOrdinal regression |
| spellingShingle | Xiaojia Zhu Rui Chen Zhiwen Shao Ming Zhang Yuhu Dai Wenzhi Zhang Chuandong Lang MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural image Alexandria Engineering Journal Computer vision Artificial intelligence Scoliosis detection Joint learning Ordinal regression |
| title | MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural image |
| title_full | MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural image |
| title_fullStr | MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural image |
| title_full_unstemmed | MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural image |
| title_short | MGScoliosis: Multi-grained scoliosis detection with joint ordinal regression from natural image |
| title_sort | mgscoliosis multi grained scoliosis detection with joint ordinal regression from natural image |
| topic | Computer vision Artificial intelligence Scoliosis detection Joint learning Ordinal regression |
| url | http://www.sciencedirect.com/science/article/pii/S1110016824012225 |
| work_keys_str_mv | AT xiaojiazhu mgscoliosismultigrainedscoliosisdetectionwithjointordinalregressionfromnaturalimage AT ruichen mgscoliosismultigrainedscoliosisdetectionwithjointordinalregressionfromnaturalimage AT zhiwenshao mgscoliosismultigrainedscoliosisdetectionwithjointordinalregressionfromnaturalimage AT mingzhang mgscoliosismultigrainedscoliosisdetectionwithjointordinalregressionfromnaturalimage AT yuhudai mgscoliosismultigrainedscoliosisdetectionwithjointordinalregressionfromnaturalimage AT wenzhizhang mgscoliosismultigrainedscoliosisdetectionwithjointordinalregressionfromnaturalimage AT chuandonglang mgscoliosismultigrainedscoliosisdetectionwithjointordinalregressionfromnaturalimage |