• Version
• Download 6
• File Size 432.87 KB
• File Count 1
• Create Date 11 June 2025
• Last Updated 11 June 2025

SMART DERMATOLOGY: SKIN DISEASE DETECTION AND RECOMMENDATION SYSTEM USING VISION TRANSFER

Authors:

1 Mr.V. UDHAYAKUMAR, 2 DEEPA SHIVANI

1 Assistant Professor, Department of Computer Applications, Sri Manakula Vinayagar Engineering College (Autonomous), Puducherry 605008, India

udhayakumar.mca@smvec.ac.in

2 Post Graduate student, Department of Computer Applications, Sri Manakula Vinayagar Engineering College (Autonomous), Puducherry 605008, India

deepashivaniaroumougam@gmail.com

ABSTRACT: Skin disorders, including acne, eczema, psoriasis and melanoma, are prevalent and can result in serious health concerns. They often represent some of the first signs of serious illnesses. Timely and accurate identification of skin disorders is paramount, however the traditional diagnostic approach is primarily visualized; therefore it is tedious, subjective and reliant on expertise. These approaches tend to miss the subtle differences between diseases that have a similar visual presentation, in addition, they lack personalized value propositions for caring for patients. In this project we are proposing a next generation diagnostic solution in which the classifications utilize a hybrid of Vision Transformers (ViT) and Convolutional Neural Networks (CNNs). The ViT model partitions the input images into patches and then some self-attention module is used to learn the global context as well the subtle features of the lesion. The CNN modules will use the ViT-processed feature to provide overall classification of skin diseases including melanoma, basal cell carcinoma, and benign keratosis in addition to being able to provide care options. The proposed system will suggest treatments, check physician availability to dermatologists or specialized hospitals nearby, and suggest a remedy to take responsibly at home based on the diagnosis. The proposed solution will employ a recommendation system to connect the deep learning to the skin disease function with the system that is expanding the diagnostic accuracy, provide results significantly faster and egregiously improve the process accessibility of the progressively serious illness. It is a potent and scalable model that gives us a chance to overhaul the way skin disease diagnostics is done, while also empowering patients by better, more timely and higher-quality medical intervention possibly resulting in better patient outcomes.

Keywords: Skin Disease Detection, Vision Transformer (ViT), Convolutional Neural Network (CNN), Deep Learning, Self-Attention Mechanism, Image-Based Diagnosis, Basal Cell Carcinoma.