A Novel Ensemble CNN Model for Accurate Detection of Melanoma

Melanoma is a type of skin cancer that develops from melanocytes, the cells responsible for the production of the pigment melanin. If not detected early, melanoma can be life-threatening as it has a high potential to spread to other organs, making treatment more challenging and decreasing survival rates. Artificial intelligence (AI) speeds up the detection of melanoma by rapidly analyzing images, overcoming the time-consuming limitations of traditional methods for earlier predictions. This study employs a deep learning framework to identify melanoma using dermoscopic images. We proposed a novel ensemble model that combines two convolutional neural networks (CNNs), DenseNet and Inception. Our approach achieved an accuracy of 94.59% in the test set while utilizing the publicly available melanoma dataset from Kaggle.