Related Experiment Video
Updated: May 21, 2026
07:51
Hydra, a Computer-Based Platform for Aiding Clinicians in Cardiovascular Analysis and Diagnosis
Published on: September 26, 2018
HyRA-CXR: a hybrid residual-attention deep network for chest X-ray classification
Ahmeed Suliman Farhan1, Umar Manzoor2, Ali Al-Kubaisi3
1Electronic Computer Center, University of Anbar, Ramadi, Iraq.
Frontiers in Artificial Intelligence
|May 20, 2026
Summary
A new deep learning model, HyRA-CXR, enhances chest X-ray (CXR) classification accuracy for pulmonary diseases. This hybrid residual-attention network offers a compact and efficient solution for improved diagnostic speed and reliability.
Area of Science:
- Artificial Intelligence
- Medical Imaging
- Deep Learning
Background:
- Chest X-ray (CXR) interpretation is critical for diagnosing pulmonary diseases but faces challenges with manual reading speed and accuracy.
- High-volume settings and resource limitations exacerbate diagnostic delays and potential errors.
Purpose of the Study:
- To introduce HyRA-CXR, a novel hybrid residual-attention convolutional neural network for automated CXR classification.
- To improve diagnostic efficiency and accuracy in pulmonary disease detection.
Main Methods:
- Developed a hybrid residual-attention convolutional neural network (HyRA-CXR) integrating residual blocks and dual attention mechanisms.
- Optimized hyperparameters using KerasTuner and evaluated the model via five-fold stratified cross-validation on the Lung X-Ray Image dataset.
Main Results:
- HyRA-CXR achieved an average accuracy of 90.39%, surpassing DenseNet121 (89.38%) and Xception (89.12%).
- Experimental results confirmed the significant contribution of both residual and attention modules to model performance.
- The model maintains a compact architecture with 0.52M parameters.
Conclusions:
- HyRA-CXR demonstrates competitive accuracy and efficiency for automated CXR classification.
- The model's compact architecture makes it suitable for deployment in resource-constrained environments.
- The study provides a promising AI-driven tool to enhance pulmonary disease diagnosis.