Hybrid Convolutional Neural Network Architectures for Skin Cancer Classification

Skin cancer is a common form of seen in humans. Like other diseases, early diagnosis skin vital. In the study, deep learning architectures, which are popular machine algorithms, used to classify cancer. order increase accuracy performance, hybrid structures realized using K-Nearest neighbor (KNN), Support vector (SVM) and Decision tree (DT). After feature extraction convolutional neural network, KNN, SVM DT applied separately for classification. While KNN produced use decision has negatively affected performance. training validation processes with seven-class mnist: ham10000 dataset containing dermatological images, confusion matrix criteria architectures reported. Eight different implemented. The highest provided by structure last layer Alexnet architecture replaced classifier.