M-GRADCAM-Explainable MobileNetV2 Framework for detection of Papulosquamous Skin Diseases Using Grad-CAM and CLAHE

Fast growth of the Papulosquamous skin diseases, such as Psoriasis, Lichen Planus, Pityriasis Rosea, and Pityriasis Rubra Pilaris, accurate detection of Papulosquamous skin diseases, from healing imageries is necessary. It poses significant diagnostic challenges due to their clinical and morphological similarities. It is critical for timely diagnosis and treatment. This paper presents an innovative approach that research integrates explainable AI (XAI) techniques, specifically Gradient-weighted Class Activation Mapping (Grad-CAM)) for highlight critical regions(such as texture, color variations, shape, and lesion distribution patterns) influencing the classification ensuring a transparent and explainable decision making process for image visualization and CALCHE used for balanced the imbalance dataset with an optimized MobileNetV2 approach to achieve efficient and high accuracy classification of Papulosquamous skin diseases. The proposed technique decreases computational intricacy while continuing evaluation performance. The proposed MobilenetV2 model is evaluated using precision, recall, F1-score, and AUC-ROC, achieving an outstanding classification accuracy of 96%, outperforming traditional classifiers like KNN, RF, ANN, VGG16 and SVM. The results highlight MobileNetV2 superior discriminatory capability in identifying visually similar dermatological conditions. This study includes design the user interface for the Papulosquamous Detection App, where the MobilenetV2 model accurately detects Papulosquamous skin disorder in the uploaded image and displays the confidence level of the detection.