This project focuses on improving deep learning models for healthcare by enhancing their transparency and reasoning. Although these models are effective in tasks like detection, classification, and segmentation, their decision-making processes are often unclear, which can reduce trust among healthcare professionals. We aim to develop AI-driven healthcare solutions that offer both high accuracy and clear insights into how decisions are made. By advancing the performance and interpretability of these models, we intend to build greater confidence in AI tools and support their effective application in clinical environments.