Automatic Detection and Segmentation of Melanoma using Fuzzy C-Means

Due to the low contrast between lesion and the surrounding skin, the automatic segmentation of skin lesions in dermoscopy images is a difficult task. For human machine interaction applications, a good skin detector capable of capturing skin tones under various conditions is important. Skin detectors such as skin probability maps or Gaussian mixture models generally achieve acceptable results in skin segmentation. The mortality rate of melanoma is the highest among patients with skin cancer. Melanoma threatens life when it grows beyond the skin dermis. Depth is therefore an important factor in the diagnosis of melanoma. This paper introduces a computerized non - invasive dermoscopy system that takes into account the estimated depth of skin lesions for diagnosis. A new computer - aided diagnostic system (CAD) is proposed for the detection of malignant melanoma from dermoscopy images based on Fuzzy C means. The proposed simulation measures the accurate diagnosis of the image of Ground Truth and the segmented image and confirms the accuracy values of up to 99.21% for DermQuest dataset, 96.05% for DermIS Dataset and 96.48% for ISIC 2016 dataset.