Dynamic Channel Attention for Enhanced Spatial Feature Extraction in Medical Image Analysis using Advanced Attention Capsule Network

Medical image analysis is essential in healthcare, guiding diagnosis, treatment, and monitoring. This study presents AACNet (Advanced Attention Capsule Network), a deep learning framework addressing the complexity of diverse medical images. AACNet incorporates a multi-feature extractor with SPP layer, multi-level capsule network, and dynamic channel attention modules. Trained on curated datasets, including chest X-rays and CT scans, augmented for enhanced generalization, AACNet achieves 92.43% accuracy on X-rays and 94.64% on CT scans, surpassing other models in multiple metrics. The model's interpretability, utilizing dynamic channel attention, underscores its capacity to emphasize crucial spatial features. The innovative integration of dynamic channel attention and capsule networks makes AACNet a pivotal solution for medical image analysis. The research findings underscore the model's adaptability, effectiveness, and interpretability. AACNet emerges as a pivotal solution for medical image analysis, exhibiting consistent superior performance and potential for real-world clinical applications.