A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injury
Introduction: Precise classification has an important role in treatment of pressure injury (PI), while current machine-learning or deeplearning based methods of PI classification remain low accuracy.Methods: In this study, we developed a deeplearning based weighted feature fusion architecture for fi...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2024-02-01
|
| Series: | Frontiers in Physiology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fphys.2024.1304829/full |
| _version_ | 1797300284135833600 |
|---|---|
| author | Dongfang Wang Dongfang Wang Lirui Guo Juan Zhong Huodan Yu Yadi Tang Li Peng Qiuni Cai Yangzhi Qi Dong Zhang Puxuan Lin |
| author_facet | Dongfang Wang Dongfang Wang Lirui Guo Juan Zhong Huodan Yu Yadi Tang Li Peng Qiuni Cai Yangzhi Qi Dong Zhang Puxuan Lin |
| author_sort | Dongfang Wang |
| collection | DOAJ |
| description | Introduction: Precise classification has an important role in treatment of pressure injury (PI), while current machine-learning or deeplearning based methods of PI classification remain low accuracy.Methods: In this study, we developed a deeplearning based weighted feature fusion architecture for fine-grained classification, which combines a top-down and bottom-up pathway to fuse high-level semantic information and low-level detail representation. We validated it in our established database that consist of 1,519 images from multi-center clinical cohorts. ResNeXt was set as the backbone network.Results: We increased the accuracy of stage 3 PI from 60.3% to 76.2% by adding weighted feature pyramid network (wFPN). The accuracy for stage 1, 2, 4 PI were 0.870, 0.788, and 0.845 respectively. We found the overall accuracy, precision, recall, and F1-score of our network were 0.815, 0.808, 0.816, and 0.811 respectively. The area under the receiver operating characteristic curve was 0.940.Conclusions: Compared with current reported study, our network significantly increased the overall accuracy from 75% to 81.5% and showed great performance in predicting each stage. Upon further validation, our study will pave the path to the clinical application of our network in PI management. |
| first_indexed | 2024-03-07T23:05:01Z |
| format | Article |
| id | doaj.art-4c2e298c19a747c09cdf88bfcbfee377 |
| institution | Directory Open Access Journal |
| issn | 1664-042X |
| language | English |
| last_indexed | 2024-03-07T23:05:01Z |
| publishDate | 2024-02-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Physiology |
| spelling | doaj.art-4c2e298c19a747c09cdf88bfcbfee3772024-02-22T05:19:19ZengFrontiers Media S.A.Frontiers in Physiology1664-042X2024-02-011510.3389/fphys.2024.13048291304829A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injuryDongfang Wang0Dongfang Wang1Lirui Guo2Juan Zhong3Huodan Yu4Yadi Tang5Li Peng6Qiuni Cai7Yangzhi Qi8Dong Zhang9Puxuan Lin10Department of Neurosurgery, Wuhan University Renmin Hospital, Wuhan, ChinaSchool of Physics and Technology, Wuhan University, Wuhan, ChinaDepartment of Neurosurgery, Wuhan University Renmin Hospital, Wuhan, ChinaDepartment of Neurosurgery, Wuhan University Renmin Hospital, Wuhan, ChinaDepartment of Neurosurgery, Wuhan University Renmin Hospital, Wuhan, ChinaDepartment of Neurosurgery, Wuhan University Renmin Hospital, Wuhan, ChinaUnion Hospital Tongji Medical College, Huazhong University of Science and Technology, Wuhan, ChinaNeurosurgery Department, Zhongshan Hospital Xiamen University, Xiamen, ChinaDepartment of Neurosurgery, Wuhan University Renmin Hospital, Wuhan, ChinaSchool of Physics and Technology, Wuhan University, Wuhan, ChinaDepartment of Neurosurgery, Wuhan University Renmin Hospital, Wuhan, ChinaIntroduction: Precise classification has an important role in treatment of pressure injury (PI), while current machine-learning or deeplearning based methods of PI classification remain low accuracy.Methods: In this study, we developed a deeplearning based weighted feature fusion architecture for fine-grained classification, which combines a top-down and bottom-up pathway to fuse high-level semantic information and low-level detail representation. We validated it in our established database that consist of 1,519 images from multi-center clinical cohorts. ResNeXt was set as the backbone network.Results: We increased the accuracy of stage 3 PI from 60.3% to 76.2% by adding weighted feature pyramid network (wFPN). The accuracy for stage 1, 2, 4 PI were 0.870, 0.788, and 0.845 respectively. We found the overall accuracy, precision, recall, and F1-score of our network were 0.815, 0.808, 0.816, and 0.811 respectively. The area under the receiver operating characteristic curve was 0.940.Conclusions: Compared with current reported study, our network significantly increased the overall accuracy from 75% to 81.5% and showed great performance in predicting each stage. Upon further validation, our study will pave the path to the clinical application of our network in PI management.https://www.frontiersin.org/articles/10.3389/fphys.2024.1304829/fullpressure injurydeep-learningfine-grained classificationweighted feature pyramid networkclassification |
| spellingShingle | Dongfang Wang Dongfang Wang Lirui Guo Juan Zhong Huodan Yu Yadi Tang Li Peng Qiuni Cai Yangzhi Qi Dong Zhang Puxuan Lin A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injury Frontiers in Physiology pressure injury deep-learning fine-grained classification weighted feature pyramid network classification |
| title | A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injury |
| title_full | A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injury |
| title_fullStr | A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injury |
| title_full_unstemmed | A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injury |
| title_short | A novel deep-learning based weighted feature fusion architecture for precise classification of pressure injury |
| title_sort | novel deep learning based weighted feature fusion architecture for precise classification of pressure injury |
| topic | pressure injury deep-learning fine-grained classification weighted feature pyramid network classification |
| url | https://www.frontiersin.org/articles/10.3389/fphys.2024.1304829/full |
| work_keys_str_mv | AT dongfangwang anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT dongfangwang anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT liruiguo anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT juanzhong anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT huodanyu anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT yaditang anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT lipeng anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT qiunicai anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT yangzhiqi anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT dongzhang anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT puxuanlin anoveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT dongfangwang noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT dongfangwang noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT liruiguo noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT juanzhong noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT huodanyu noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT yaditang noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT lipeng noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT qiunicai noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT yangzhiqi noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT dongzhang noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury AT puxuanlin noveldeeplearningbasedweightedfeaturefusionarchitectureforpreciseclassificationofpressureinjury |