A Comparative Evaluation of Machine Learning based Approaches for Detection of Papulosquamous Skin Diseases

Papulosquamous skin diseases are challenging to diagnose due to overlapping features like erythema and scaling. Traditional Machine Learning models demonstrated moderate effectiveness but lagged in accuracy compared to Deep Learning models, which excel in processing complex dermatological images. This study highlights the value of diverse, highquality datasets in advancing diagnostics for rare Papulosquamous skin conditions. This paper evaluates importance of ML (Machine Learning) and DL(deep learning) models used to classifying these Papulosquamous skin disorders using dermatological image datasets. Evaluation performance criteria including accuracy, precision, and recall are used to compare three Deep Learning models (ResNet50, VGG16, and InceptionV3) and six Machine Learning classifiers (SVM, LR, RF, DT, XGBoost and ANN). InceptionV3 achieved with highest accuracy (92%), followed by ResNet50 and VGG16. The results highlight the potential of DL models is improving diagnostic accuracy for rare skin diseases.