A proposed architecture for convolutional neural networks to detect skin cancers

The goal of the research paper is to design and development a computer-based system for segmentation classification malignant skin diseases comparison between accuracy their detection, as two were detected. Namely, basal cell carcinoma melanoma separately with images nevus, collected from ISIC 2020 archive group, total, used: 17,846 include 3,008 (BCC), 5,272 melanoma, 9,566 validation data contains 20% used which are not classified randomly taken set images, final test 1,500 anonymous images. An architecture convolutional neural network technology in deep learning has been proposed that consists layers processing. Processing raw input group pre-treatment transformations, augmentation process, so number became 86094 27,072 BCC, 47,448 melanoma. Through detection BCC was 98.25%, higher than 91.61%.