Early and accurate diagnosis of multiple diseases using medical imaging is critical for effective healthcare delivery. This paper proposes a generalized deep learning framework for simultaneous detection of bone fractures, brain tumors, lung cancer, and skin lesions using heterogeneous medical image datasets. A lightweight Convolutional Neural Network (CNN) is designed to extract hierarchical features and perform multi-class classification. The model achieves near-perfect accuracy (100%), with excellent precision, recall, and F1-scores across all classes. ROC-AUC scores exceeding 0.99 further validate the model’s discriminative capability. Despite strong performance, cross-validation results highlight the need for improved generalization. The proposed system demonstrates strong potential for clinical decision support applications.
A Hemanth Kumar (Wed,) studied this question.