An efficient attention Densenet with LSTM for lung disease detection and classification using X-ray images supported...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer-Aided Diagnosis

Background:

• Lung diseases pose a significant global health burden, demanding precise diagnostic tools.
• Accurate and early detection of lung conditions is crucial for effective patient management and improved outcomes.

Purpose of the Study:

• To introduce a novel deep learning framework for enhanced lung disease classification from chest X-ray images.
• To improve the accuracy and efficiency of lung disease diagnosis through advanced image analysis techniques.

Main Methods:

• A three-phase framework involving image acquisition, segmentation, and classification using deep learning.
• Lung segmentation performed by Adaptive Recurrent Residual U-Net (AR2-UNet) optimized with Enhanced Pufferfish Optimisation Algorithm (EPOA).
• Classification utilizing an Attention-based Densenet with Long Short Term Memory (ADNet-LSTM) model.

Main Results:

• The proposed ADNet-LSTM model achieved a highest classification accuracy of 93.92%.
• Significantly outperformed baseline models: ResNet (90.77%), Inception (89.55%), DenseNet (89.66%), and LSTM (91.79%).
• Demonstrated superior performance in robust lung disease categorization.

Conclusions:

• The developed deep learning framework provides a dependable and efficient solution for lung disease detection.
• The proposed model supports clinicians in achieving earlier and more accurate diagnoses.
• Highlights the potential of advanced AI in medical diagnostics for better patient care.