A weak edge estimation based multi-task neural network for OCT segmentation.
Optical Coherence Tomography (OCT) offers high-resolution images of the eye's fundus. This enables thorough analysis of retinal health by doctors, providing a solid basis for diagnosis and treatment. With the development of deep learning, deep learning-based methods are becoming more popular fo...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2025-01-01
|
| Series: | PLoS ONE |
| Online Access: | https://doi.org/10.1371/journal.pone.0316089 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841555610100826112 |
|---|---|
| author | Fan Yang Pu Chen Shiqi Lin Tianming Zhan Xunning Hong Yunjie Chen |
| author_facet | Fan Yang Pu Chen Shiqi Lin Tianming Zhan Xunning Hong Yunjie Chen |
| author_sort | Fan Yang |
| collection | DOAJ |
| description | Optical Coherence Tomography (OCT) offers high-resolution images of the eye's fundus. This enables thorough analysis of retinal health by doctors, providing a solid basis for diagnosis and treatment. With the development of deep learning, deep learning-based methods are becoming more popular for fundus OCT image segmentation. Yet, these methods still encounter two primary challenges. Firstly, deep learning methods are sensitive to weak edges. Secondly, the high cost of annotating medical image data results in a lack of labeled data, leading to overfitting during model training. To tackle these challenges, we introduce the Multi-Task Attention Mechanism Network with Pruning (MTAMNP), consisting of a segmentation branch and a boundary regression branch. The boundary regression branch utilizes an adaptive weighted loss function derived from the Truncated Signed Distance Function(TSDF), improving the model's capacity to preserve weak edge details. The Spatial Attention Based Dual-Branch Information Fusion Block links these branches, enabling mutual benefit. Furthermore, we present a structured pruning method grounded in channel attention to decrease parameter count, mitigate overfitting, and uphold segmentation accuracy. Our method surpasses other cutting-edge segmentation networks on two widely accessible datasets, achieving Dice scores of 84.09% and 93.84% on the HCMS and Duke datasets. |
| format | Article |
| id | doaj-art-7e6f4ceee2304dcfb6efbd8d5b3356cc |
| institution | Kabale University |
| issn | 1932-6203 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Public Library of Science (PLoS) |
| record_format | Article |
| series | PLoS ONE |
| spelling | doaj-art-7e6f4ceee2304dcfb6efbd8d5b3356cc2025-01-08T05:31:44ZengPublic Library of Science (PLoS)PLoS ONE1932-62032025-01-01201e031608910.1371/journal.pone.0316089A weak edge estimation based multi-task neural network for OCT segmentation.Fan YangPu ChenShiqi LinTianming ZhanXunning HongYunjie ChenOptical Coherence Tomography (OCT) offers high-resolution images of the eye's fundus. This enables thorough analysis of retinal health by doctors, providing a solid basis for diagnosis and treatment. With the development of deep learning, deep learning-based methods are becoming more popular for fundus OCT image segmentation. Yet, these methods still encounter two primary challenges. Firstly, deep learning methods are sensitive to weak edges. Secondly, the high cost of annotating medical image data results in a lack of labeled data, leading to overfitting during model training. To tackle these challenges, we introduce the Multi-Task Attention Mechanism Network with Pruning (MTAMNP), consisting of a segmentation branch and a boundary regression branch. The boundary regression branch utilizes an adaptive weighted loss function derived from the Truncated Signed Distance Function(TSDF), improving the model's capacity to preserve weak edge details. The Spatial Attention Based Dual-Branch Information Fusion Block links these branches, enabling mutual benefit. Furthermore, we present a structured pruning method grounded in channel attention to decrease parameter count, mitigate overfitting, and uphold segmentation accuracy. Our method surpasses other cutting-edge segmentation networks on two widely accessible datasets, achieving Dice scores of 84.09% and 93.84% on the HCMS and Duke datasets.https://doi.org/10.1371/journal.pone.0316089 |
| spellingShingle | Fan Yang Pu Chen Shiqi Lin Tianming Zhan Xunning Hong Yunjie Chen A weak edge estimation based multi-task neural network for OCT segmentation. PLoS ONE |
| title | A weak edge estimation based multi-task neural network for OCT segmentation. |
| title_full | A weak edge estimation based multi-task neural network for OCT segmentation. |
| title_fullStr | A weak edge estimation based multi-task neural network for OCT segmentation. |
| title_full_unstemmed | A weak edge estimation based multi-task neural network for OCT segmentation. |
| title_short | A weak edge estimation based multi-task neural network for OCT segmentation. |
| title_sort | weak edge estimation based multi task neural network for oct segmentation |
| url | https://doi.org/10.1371/journal.pone.0316089 |
| work_keys_str_mv | AT fanyang aweakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT puchen aweakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT shiqilin aweakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT tianmingzhan aweakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT xunninghong aweakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT yunjiechen aweakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT fanyang weakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT puchen weakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT shiqilin weakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT tianmingzhan weakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT xunninghong weakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation AT yunjiechen weakedgeestimationbasedmultitaskneuralnetworkforoctsegmentation |