Melanoma Skin Cancer Prediction Using VGG۱۹ and DenseNet۱۲۱

Melanoma skin cancer is one of the most lethal forms of cancer. Deep learning has shown promising results in predicting melanoma skin cancer. This paper proposes a novel method for skin cancer prediction using two modified networks: DenseNet۱۲۱ and VGG۱۹. In the proposed method, image augmentation is initially performed to increase the diversity of the training data, thereby enhancing the model’s generalization capability. Subsequently, the capabilities of the two networks are enhanced by adding two dense layers with ۲۵۶۰ and ۱۲۰۰ neurons, respectively. The proposed method is applied to the three-class ISIC-۲۰۱۷ dataset, taking into account metrics such as Accuracy, Precision, Recall, and F-measures. The results of the proposed method have shown improvement over many other published works, demonstrating its effectiveness in skin cancer prediction.

کلیدواژه ها:

Deep learning ، DenseNet ، Melanoma skin cancer ، VGG۱۹