High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images
The development of automatic chest X-ray (CXR) disease classification algorithms is significant for diagnosing thoracic diseases. Owing to the characteristics of lesions in CXR images, including high similarity in appearance of the disease, varied sizes, and different occurrence locations, most exis...
Autores principales: | , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
MDPI
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10340191/ https://www.ncbi.nlm.nih.gov/pubmed/37443559 http://dx.doi.org/10.3390/diagnostics13132165 |
_version_ | 1785072019609485312 |
---|---|
author | Li, Qiang Chen, Mingyu Geng, Jingjing Adamu, Mohammed Jajere Guan, Xin |
author_facet | Li, Qiang Chen, Mingyu Geng, Jingjing Adamu, Mohammed Jajere Guan, Xin |
author_sort | Li, Qiang |
collection | PubMed |
description | The development of automatic chest X-ray (CXR) disease classification algorithms is significant for diagnosing thoracic diseases. Owing to the characteristics of lesions in CXR images, including high similarity in appearance of the disease, varied sizes, and different occurrence locations, most existing convolutional neural network-based methods have insufficient feature extraction for thoracic lesions and struggle to adapt to changes in lesion size and location. To address these issues, this study proposes a high-resolution classification network with dynamic convolution and coordinate attention (HRCC-Net). In the method, this study suggests a parallel multi-resolution network in which a high-resolution branch acquires essential detailed features of the lesion and multi-resolution feature swapping and fusion to obtain multiple receptive fields to extract complicated disease features adequately. Furthermore, this study proposes dynamic convolution to enhance the network’s ability to represent multi-scale information to accommodate lesions of diverse scales. In addition, this study introduces a coordinate attention mechanism, which enables automatic focus on pathologically relevant regions and capturing the variations in lesion location. The proposed method is evaluated on ChestX-ray14 and CheXpert datasets. The average AUC (area under ROC curve) values reach 0.845 and 0.913, respectively, indicating this method’s advantages compared with the currently available methods. Meanwhile, with its specificity and sensitivity to measure the performance of medical diagnostic systems, the network can improve diagnostic efficiency while reducing the rate of misdiagnosis. The proposed algorithm has great potential for thoracic disease diagnosis and treatment. |
format | Online Article Text |
id | pubmed-10340191 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2023 |
publisher | MDPI |
record_format | MEDLINE/PubMed |
spelling | pubmed-103401912023-07-14 High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images Li, Qiang Chen, Mingyu Geng, Jingjing Adamu, Mohammed Jajere Guan, Xin Diagnostics (Basel) Article The development of automatic chest X-ray (CXR) disease classification algorithms is significant for diagnosing thoracic diseases. Owing to the characteristics of lesions in CXR images, including high similarity in appearance of the disease, varied sizes, and different occurrence locations, most existing convolutional neural network-based methods have insufficient feature extraction for thoracic lesions and struggle to adapt to changes in lesion size and location. To address these issues, this study proposes a high-resolution classification network with dynamic convolution and coordinate attention (HRCC-Net). In the method, this study suggests a parallel multi-resolution network in which a high-resolution branch acquires essential detailed features of the lesion and multi-resolution feature swapping and fusion to obtain multiple receptive fields to extract complicated disease features adequately. Furthermore, this study proposes dynamic convolution to enhance the network’s ability to represent multi-scale information to accommodate lesions of diverse scales. In addition, this study introduces a coordinate attention mechanism, which enables automatic focus on pathologically relevant regions and capturing the variations in lesion location. The proposed method is evaluated on ChestX-ray14 and CheXpert datasets. The average AUC (area under ROC curve) values reach 0.845 and 0.913, respectively, indicating this method’s advantages compared with the currently available methods. Meanwhile, with its specificity and sensitivity to measure the performance of medical diagnostic systems, the network can improve diagnostic efficiency while reducing the rate of misdiagnosis. The proposed algorithm has great potential for thoracic disease diagnosis and treatment. MDPI 2023-06-25 /pmc/articles/PMC10340191/ /pubmed/37443559 http://dx.doi.org/10.3390/diagnostics13132165 Text en © 2023 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 Li, Qiang Chen, Mingyu Geng, Jingjing Adamu, Mohammed Jajere Guan, Xin High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images |
title | High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images |
title_full | High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images |
title_fullStr | High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images |
title_full_unstemmed | High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images |
title_short | High-Resolution Network with Dynamic Convolution and Coordinate Attention for Classification of Chest X-ray Images |
title_sort | high-resolution network with dynamic convolution and coordinate attention for classification of chest x-ray images |
topic | Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10340191/ https://www.ncbi.nlm.nih.gov/pubmed/37443559 http://dx.doi.org/10.3390/diagnostics13132165 |
work_keys_str_mv | AT liqiang highresolutionnetworkwithdynamicconvolutionandcoordinateattentionforclassificationofchestxrayimages AT chenmingyu highresolutionnetworkwithdynamicconvolutionandcoordinateattentionforclassificationofchestxrayimages AT gengjingjing highresolutionnetworkwithdynamicconvolutionandcoordinateattentionforclassificationofchestxrayimages AT adamumohammedjajere highresolutionnetworkwithdynamicconvolutionandcoordinateattentionforclassificationofchestxrayimages AT guanxin highresolutionnetworkwithdynamicconvolutionandcoordinateattentionforclassificationofchestxrayimages |