Enhancing Skin Cancer Detection with a Modified Deep CNN Classifier

Skin cancer, a potentially fatal condition arising from abnormal skin cell growth due to ultraviolet exposure, demands early detection to reduce mortality rates. Despite advancements in diagnostic approaches, challenges such as enhancing accuracy and minimizing computational complexity remain. In order to reduce mistakes, boost robustness, and quicken convergence, this study presents an upgraded deep convolutional neural system (CNN) coder that has been fine-tuned using a sophisticated tuning method. The model demonstrates superior performance in detecting skin cancer by effectively analyzing critical disease features. Experimental results highlight its efficiency, achieving 96.52% accuracy, 96.69% sensitivity, and 96.54% specificity during training and 93.59%, 92.14%, and 95.22% respectively in k-fold cross-validation. These findings emphasize the model’s potential to outperform existing approaches, offering a reliable and efficient solution for skin cancer diagnosis.