Melanoma Diagnosis System using Artificial Neural Networks


  • Raafat K. Oubida


Skin cancer, Biopsy, Segmentation, ABCD rule.


In today's era, melanoma is one of the most precarious and deadliest of any other type of skin cancer. Both Benign and Malignant Melanoma, if detection of malignant melanoma at the earliest stage could be reducing the mortality rate. Many diagnostic existing systems play important role in early detection of malignant melanoma and provide the opportunity to helpful dermatologists in the classification of features from skin lesion images, which are not visible to the naked eye. This paper is intended to introduce a system that incorporates digital image processing approaches and Artificial Neural Networks (ANN) to enhance the successful of survival rate and considerably to reduce the efforts and time consumption of dermatologists and patients utilizes for the classification of Malignant Melanoma (MM) early stage. The diagnosing methodology concentrates on Fuzzy C-Means (FCM) to extract skin lesions in clinical images, subsequently feed forward neural networks to classifier of malignant melanoma vs. benign melanoma on real melanoma images. Experimental results illustrates that the fuzzification of feature modelling provide good results in term of sensitivity (74.6%) with the specificity 76.2% and classification accuracy of (75.3%).  The system can be used as an effective an inexpensive tool to assist dermatologist for diagnosis accuracy as well as the time consuming and costs.



How to Cite

Raafat K. Oubida. (2022). Melanoma Diagnosis System using Artificial Neural Networks. Yantu Gongcheng Xuebao/Chinese Journal of Geotechnical Engineering, 44(12), 1–8. Retrieved from