SkinAI: Affordable and Accessible Skin Cancer Diagnosis with Deep Learning and HAM10000

Skin cancer is one of the world's most common cancers, and chances of a good prognosis are significantly increased by early diagnosis. However, the underprivileged sections of society lack access to early and affordable diagnostic facilities. We fill this void by introducing SkinAI, a web-based deep learning model that employs the Xception convolutional neural network to classify dermoscopic images of skin lesions into seven diagnostic categories. For class balancing, the model is trained on the HAM10000 dataset with robust pre- processing techniques like image augmentation and SMOTE. The system includes downloadable PDF reports, real-time prediction, and a user-friendly interface which was developed with Flask using TailwindCSS. SkinAI is a user-friendly diagnostic tool that enables users to make well-informed healthcare decisions by providing downloadable reports and real-time predictions, especially in regions with limited dermatological care accessibility.