AI-Powered Skin Cancer Screening Integrating Image Processing with Deep Learning for Enhanced Accuracy

This project focuses on the development of a skin cancer detection system using image processing techniques and deep learning algorithms.A skin cancer dataset was obtained from the Kaggle website and processed through several stages to enhance image quality and extract relevant features.Initially, input images were resized to 299x299 pixels, followed by noise reduction using median filtering across the RGB color channels.Contrast adjustment was performed to enhance image details, which were then converted to grayscale for edge detection using the Canny method.The edges were dilated to close gaps, followed by computing the distance transform and applying watershed segmentation to identify distinct regions.The processed images were then classified using a modified InceptionV3 model, where the network was fine-tuned for five different skin conditions, including melanoma and basal cell carcinoma.The model was trained and validated, achieving notable accuracy.Furthermore, performance metrics such as precision, recall, and F1 score were calculated to evaluate the classification effectiveness.The system also provided treatment recommendations based on the predicted skin conditions, highlighting the potential of integrating advanced image processing and deep learning techniques for effective skin cancer diagnosis.This study demonstrates a significant step toward developing an automated skin cancer detection tool that can aid medical professionals in early diagnosis and treatment planning, ultimately improving patient outcomes in dermatology.