Publication

Abstract : Leprosy remains a significant public health challenge, particularly in India. Early and accurate diagnosis is crucial for effective treatment and preventing nerve damage. This study explores the potential of deep learning techniques to improve leprosy diagnosis, specifically focusing on tuber-culoid leprosy. Convolutional Neural Networks (CNNs) were investigated for leprosy detection using skin lesion images. A comparative analysis was conducted using VGG16, MobileNet v1, Xception v1 and EfficientNet B0 models. With a recall of 0.9298, an F1 score of 0.9217, and an accuracy of 0.9138, EfficientNetB0 performed the best. This indicates that EfficientNetB0 is suitable for leprosy diagnosis due to its ability to balance precision and recall. The study highlights the potential of deep learning for improving leprosy diagnosis and paves the way for more accurate, reliable, and accessible diagnostic tools.