Automated skin lesion classification using deep custom convolutional neural network

Skin cancer is the most prevalent cancer in the world, and early diagnosis is the key to successful treatment. Conventional diagnosis is time-consuming and dependent on skilled dermatologists. In this project, we suggest a Convolutional Neural Network (CNN)-based approach for automatic skin cancer detection from dermoscopic images. The system employs preprocessed images from pretrained deep learning models and expert CNN architectures for the classification of skin lesions as benign or malignant. Image preprocessing, data augmentation, and transfer learning are implemented for improving accuracy. The model is tested with some metrics like accuracy, precision, recall, and F1-score. The AI-based system is suggested to improve early diagnosis, remove human error, and assist healthcare professionals in making faster, better decisions