Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural network
Abstract To address misdiagnosis caused by feature coupling in multi-label medical image classification, this study introduces a chest X-ray pathology reasoning method. It combines hierarchical attention convolutional networks with a multi-label decoupling loss function. This method aims to enhance...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-07-01
|
| Series: | BMC Medical Imaging |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12880-025-01800-3 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849238193405165568 |
|---|---|
| author | Shouyi Yang Yongxin Wu |
| author_facet | Shouyi Yang Yongxin Wu |
| author_sort | Shouyi Yang |
| collection | DOAJ |
| description | Abstract To address misdiagnosis caused by feature coupling in multi-label medical image classification, this study introduces a chest X-ray pathology reasoning method. It combines hierarchical attention convolutional networks with a multi-label decoupling loss function. This method aims to enhance the precise identification of complex lesions. It dynamically captures multi-scale lesion morphological features and integrates lung field partitioning with lesion localization through a dual-path attention mechanism, thereby improving clinical disease prediction accuracy. An adaptive dilated convolution module with 3 × 3 deformable kernels dynamically captures multi-scale lesion features. A channel-space dual-path attention mechanism enables precise feature selection for lung field partitioning and lesion localization. Cross-scale skip connections fuse shallow texture and deep semantic information, enhancing microlesion detection. A KL divergence-constrained contrastive loss function decouples 14 pathological feature representations via orthogonal regularization, effectively resolving multi-label coupling. Experiments on ChestX-ray14 show a weighted F1-score of 0.97, Hamming Loss of 0.086, and AUC values exceeding 0.94 for all pathologies. This study provides a reliable tool for multi-disease collaborative diagnosis. |
| format | Article |
| id | doaj-art-fd04f36aed554d56ad50d8ae079e8dc3 |
| institution | Kabale University |
| issn | 1471-2342 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Medical Imaging |
| spelling | doaj-art-fd04f36aed554d56ad50d8ae079e8dc32025-08-20T04:01:43ZengBMCBMC Medical Imaging1471-23422025-07-0125111710.1186/s12880-025-01800-3Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural networkShouyi Yang0Yongxin Wu1Department of Medical School, Kunming University of Science and TechnologyDepartment of Cardiology, The First People’s Hospital of Yunnan ProvinceAbstract To address misdiagnosis caused by feature coupling in multi-label medical image classification, this study introduces a chest X-ray pathology reasoning method. It combines hierarchical attention convolutional networks with a multi-label decoupling loss function. This method aims to enhance the precise identification of complex lesions. It dynamically captures multi-scale lesion morphological features and integrates lung field partitioning with lesion localization through a dual-path attention mechanism, thereby improving clinical disease prediction accuracy. An adaptive dilated convolution module with 3 × 3 deformable kernels dynamically captures multi-scale lesion features. A channel-space dual-path attention mechanism enables precise feature selection for lung field partitioning and lesion localization. Cross-scale skip connections fuse shallow texture and deep semantic information, enhancing microlesion detection. A KL divergence-constrained contrastive loss function decouples 14 pathological feature representations via orthogonal regularization, effectively resolving multi-label coupling. Experiments on ChestX-ray14 show a weighted F1-score of 0.97, Hamming Loss of 0.086, and AUC values exceeding 0.94 for all pathologies. This study provides a reliable tool for multi-disease collaborative diagnosis.https://doi.org/10.1186/s12880-025-01800-3Medical image analysisConvolutional neural networkHierarchical attention mechanismMulti-label decouplingChest X-ray imaging |
| spellingShingle | Shouyi Yang Yongxin Wu Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural network BMC Medical Imaging Medical image analysis Convolutional neural network Hierarchical attention mechanism Multi-label decoupling Chest X-ray imaging |
| title | Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural network |
| title_full | Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural network |
| title_fullStr | Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural network |
| title_full_unstemmed | Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural network |
| title_short | Intelligent diagnosis model for chest X-ray images diseases based on convolutional neural network |
| title_sort | intelligent diagnosis model for chest x ray images diseases based on convolutional neural network |
| topic | Medical image analysis Convolutional neural network Hierarchical attention mechanism Multi-label decoupling Chest X-ray imaging |
| url | https://doi.org/10.1186/s12880-025-01800-3 |
| work_keys_str_mv | AT shouyiyang intelligentdiagnosismodelforchestxrayimagesdiseasesbasedonconvolutionalneuralnetwork AT yongxinwu intelligentdiagnosismodelforchestxrayimagesdiseasesbasedonconvolutionalneuralnetwork |