Enhancing Skin Cancer Detection: A Comparative Study of Meta-Heuristic Optimized Convolutional Neural Networks

CNNs are transforming into most common in a variety of fields, including medical imaging,and are generally acknowledged as an effective tool in the deep learning space. A critical usecase in this area is supporting professionals in the early detection of skin cancer bydermoscopy, which lowers death rates. Nonetheless, a number of variables may affect howaccurate system diagnostics are. Researchers have recently found computer-aided technologyto be an interesting field of study. In order to effectively diagnose skin cancer photos, thisresearch uses a state-of-the-art convolutional neural network (CNN) classifier that was tunedusing meta-heuristic approaches. Specifically built network models for visual datasets areused to train the classifier. There are several techniques to enhance the productivity of theneural network learning process. However, there is a dearth of research on the practical usesof deep learning-based neural networks. This study focuses on optimizing the weights andbiases in convolutional neural network (CNN) models by employing a novel approach thatutilizes the whale optimization algorithm. The effectiveness of this technique is evaluatedusing two skin cancer datasets, namely the DermQuest Database and the University ofWaterloo skin cancer database Dataset.