Divide-and-Attention Network for HE-Stained Pathological Image Classification
Since pathological images have some distinct characteristics that are different from natural images, the direct application of a general convolutional neural network cannot achieve good classification performance, especially for fine-grained classification problems (such as pathological image gradin...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-06-01
|
| Series: | Biology |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-7737/11/7/982 |
| _version_ | 1827597933942407168 |
|---|---|
| author | Rui Yan Zhidong Yang Jintao Li Chunhou Zheng Fa Zhang |
| author_facet | Rui Yan Zhidong Yang Jintao Li Chunhou Zheng Fa Zhang |
| author_sort | Rui Yan |
| collection | DOAJ |
| description | Since pathological images have some distinct characteristics that are different from natural images, the direct application of a general convolutional neural network cannot achieve good classification performance, especially for fine-grained classification problems (such as pathological image grading). Inspired by the clinical experience that decomposing a pathological image into different components is beneficial for diagnosis, in this paper, we propose a <b>D</b>ivide-and-<b>A</b>ttention <b>Net</b>work (<b>DANet</b>) for Hematoxylin-and-Eosin (HE)-stained pathological image classification. The DANet utilizes a deep-learning method to decompose a pathological image into nuclei and non-nuclei parts. With such decomposed pathological images, the DANet first performs feature learning independently in each branch, and then focuses on the most important feature representation through the branch selection attention module. In this way, the DANet can learn representative features with respect to different tissue structures and adaptively focus on the most important ones, thereby improving classification performance. In addition, we introduce deep canonical correlation analysis (DCCA) constraints in the feature fusion process of different branches. The DCCA constraints play the role of branch fusion attention, so as to maximize the correlation of different branches and ensure that the fused branches emphasize specific tissue structures. The experimental results of three datasets demonstrate the superiority of the DANet, with an average classification accuracy of 92.5% on breast cancer classification, 95.33% on colorectal cancer grading, and 91.6% on breast cancer grading tasks. |
| first_indexed | 2024-03-09T03:41:24Z |
| format | Article |
| id | doaj.art-57713869ca0c4f8d88b10c853e490e4c |
| institution | Directory Open Access Journal |
| issn | 2079-7737 |
| language | English |
| last_indexed | 2024-03-09T03:41:24Z |
| publishDate | 2022-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Biology |
| spelling | doaj.art-57713869ca0c4f8d88b10c853e490e4c2023-12-03T14:40:23ZengMDPI AGBiology2079-77372022-06-0111798210.3390/biology11070982Divide-and-Attention Network for HE-Stained Pathological Image ClassificationRui Yan0Zhidong Yang1Jintao Li2Chunhou Zheng3Fa Zhang4High Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100045, ChinaHigh Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100045, ChinaHigh Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100045, ChinaSchool of Artificial Intelligence, Anhui University, Hefei 230093, ChinaHigh Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100045, ChinaSince pathological images have some distinct characteristics that are different from natural images, the direct application of a general convolutional neural network cannot achieve good classification performance, especially for fine-grained classification problems (such as pathological image grading). Inspired by the clinical experience that decomposing a pathological image into different components is beneficial for diagnosis, in this paper, we propose a <b>D</b>ivide-and-<b>A</b>ttention <b>Net</b>work (<b>DANet</b>) for Hematoxylin-and-Eosin (HE)-stained pathological image classification. The DANet utilizes a deep-learning method to decompose a pathological image into nuclei and non-nuclei parts. With such decomposed pathological images, the DANet first performs feature learning independently in each branch, and then focuses on the most important feature representation through the branch selection attention module. In this way, the DANet can learn representative features with respect to different tissue structures and adaptively focus on the most important ones, thereby improving classification performance. In addition, we introduce deep canonical correlation analysis (DCCA) constraints in the feature fusion process of different branches. The DCCA constraints play the role of branch fusion attention, so as to maximize the correlation of different branches and ensure that the fused branches emphasize specific tissue structures. The experimental results of three datasets demonstrate the superiority of the DANet, with an average classification accuracy of 92.5% on breast cancer classification, 95.33% on colorectal cancer grading, and 91.6% on breast cancer grading tasks.https://www.mdpi.com/2079-7737/11/7/982pathological image classificationattention mechanismconvolutional neural networkknowledge embedding |
| spellingShingle | Rui Yan Zhidong Yang Jintao Li Chunhou Zheng Fa Zhang Divide-and-Attention Network for HE-Stained Pathological Image Classification Biology pathological image classification attention mechanism convolutional neural network knowledge embedding |
| title | Divide-and-Attention Network for HE-Stained Pathological Image Classification |
| title_full | Divide-and-Attention Network for HE-Stained Pathological Image Classification |
| title_fullStr | Divide-and-Attention Network for HE-Stained Pathological Image Classification |
| title_full_unstemmed | Divide-and-Attention Network for HE-Stained Pathological Image Classification |
| title_short | Divide-and-Attention Network for HE-Stained Pathological Image Classification |
| title_sort | divide and attention network for he stained pathological image classification |
| topic | pathological image classification attention mechanism convolutional neural network knowledge embedding |
| url | https://www.mdpi.com/2079-7737/11/7/982 |
| work_keys_str_mv | AT ruiyan divideandattentionnetworkforhestainedpathologicalimageclassification AT zhidongyang divideandattentionnetworkforhestainedpathologicalimageclassification AT jintaoli divideandattentionnetworkforhestainedpathologicalimageclassification AT chunhouzheng divideandattentionnetworkforhestainedpathologicalimageclassification AT fazhang divideandattentionnetworkforhestainedpathologicalimageclassification |