Divide-and-Attention Network for HE-Stained Pathological Image Classification

SIMPLE SUMMARY: We propose a Divide-and-Attention network that can learn representative pathological image features with respect to different tissue structures and adaptively focus on the most important ones. In addition, we introduce deep canonical correlation analysis constraints in the feature fu...

Descripción completa

Detalles Bibliográficos
Autores principales: Yan, Rui, Yang, Zhidong, Li, Jintao, Zheng, Chunhou, Zhang, Fa
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9311575/
https://www.ncbi.nlm.nih.gov/pubmed/36101363
http://dx.doi.org/10.3390/biology11070982
_version_ 1784753626453901312
author Yan, Rui
Yang, Zhidong
Li, Jintao
Zheng, Chunhou
Zhang, Fa
author_facet Yan, Rui
Yang, Zhidong
Li, Jintao
Zheng, Chunhou
Zhang, Fa
author_sort Yan, Rui
collection PubMed
description SIMPLE SUMMARY: We propose a Divide-and-Attention network that can learn representative pathological image features with respect to different tissue structures and adaptively focus on the most important ones. In addition, we introduce deep canonical correlation analysis constraints in the feature fusion process of different branches, so as to maximize the correlation of different branches and ensure that the fused branches emphasize specific tissue structures. Extensive experiments on three different pathological image datasets show that the proposed method achieved competitive results. ABSTRACT: 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 Divide-and-Attention Network (DANet) 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.
format Online
Article
Text
id pubmed-9311575
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-93115752022-07-26 Divide-and-Attention Network for HE-Stained Pathological Image Classification Yan, Rui Yang, Zhidong Li, Jintao Zheng, Chunhou Zhang, Fa Biology (Basel) Article SIMPLE SUMMARY: We propose a Divide-and-Attention network that can learn representative pathological image features with respect to different tissue structures and adaptively focus on the most important ones. In addition, we introduce deep canonical correlation analysis constraints in the feature fusion process of different branches, so as to maximize the correlation of different branches and ensure that the fused branches emphasize specific tissue structures. Extensive experiments on three different pathological image datasets show that the proposed method achieved competitive results. ABSTRACT: 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 Divide-and-Attention Network (DANet) 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. MDPI 2022-06-29 /pmc/articles/PMC9311575/ /pubmed/36101363 http://dx.doi.org/10.3390/biology11070982 Text en © 2022 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Yan, Rui
Yang, Zhidong
Li, Jintao
Zheng, Chunhou
Zhang, Fa
Divide-and-Attention Network for HE-Stained Pathological Image Classification
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 Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9311575/
https://www.ncbi.nlm.nih.gov/pubmed/36101363
http://dx.doi.org/10.3390/biology11070982
work_keys_str_mv AT yanrui divideandattentionnetworkforhestainedpathologicalimageclassification
AT yangzhidong divideandattentionnetworkforhestainedpathologicalimageclassification
AT lijintao divideandattentionnetworkforhestainedpathologicalimageclassification
AT zhengchunhou divideandattentionnetworkforhestainedpathologicalimageclassification
AT zhangfa divideandattentionnetworkforhestainedpathologicalimageclassification