Diagnosis of malignant melanoma by neural network ensemble-based system utilising hand-crafted skin lesion features

Malignant melanomas are the most deadly type of skin cancer but detected early have high chances for successful treatment. In last twenty years, interest automated melanoma recognition detection and classification dynamically increased partially because public datasets appearing with dermatoscopic images lesions. Automated computer-aided in is a very challenging task due to uneven sizes, huge intra-class variation small interclass variation, existence many artifacts image. One recognized methods diagnosis ABCD method. paper, we propose an extended version this method intelligent decision support system based on neural networks that uses its results form hand-crafted features. Automatic determination features used by difficult large diversity various quality, hair, different markers other obstacles. Therefore, it was necessary apply advanced preprocessing images. The ensemble ten networks, working parallel one network using their generate final decision. This structure allowed us increase efficiency operation several percentage points compared single network. proposed trained over 5000 tested afterward 200 moles. presented can be as primary care physicians, capable self-examination dermatoscope also important tool improve biopsy making.