Hybrid Deep Learning Model for Skin Cancer Classification

Skin cancer represents a significant public health concern worldwide, with melanoma accounting for its most lethal form. Timely identification and precise categorization of skin lesions play pivotal roles in enhancing treatment efficacy and fostering better patient outcomes. Deep learning approaches have showed promise in automatically classifying skin cancer from dermatoscopic images. In this paper, propose a hybrid deep learning model for skin cancer classification, combining the strengths of VGG16 and InceptionV3 architectures. VGG16 is known for its simplicity and effectiveness in feature extraction, while InceptionV3 excels in capturing fine-grained details and global context. The proposed hybrid model leverages the complementary features of these architectures to enhance classification performance. We train the model on a dataset of dermatoscopic images, consisting of cancer types, and evaluate its performance using conventional measures such as precision, accuracy, recall, and F1-score. Our experimental outcomes reveal that the hybrid model surpasses standalone VGG16 and InceptionV3 models, achieving superior accuracy in skin cancer classification. The proposed hybrid deep learning method holds promise for improving automated skin cancer diagnosis systems and enhancing patient care in dermatology clinics.