Segmentation of skin cancer using Fuzzy U-network via deep learning

Abstract

The most common cancer in the world is skin cancer. In recent years, one of the most important challenges to public health has been melanoma, the most dangerous type of skin cancer. In this paper, a novel MFO-Fuzzy U net has been proposed to segmentation process to extract the affected area in skin cancer image.All image processing has been doen by using IoT connected by raspberry pi. The images are pre-processed using bilateral filter for eliminating the irrelevant noise artifacts. The pre-processed images are taken as input to Fuzzy U-net for overlapped on the skin region. Consequently, the images are fed as input to Fuzzy U-net for segmentation and optimized by May Fly Optimizer is improve the range of accuracy.The performance of the existing method was compared using accuracy, precision, specificity, recall and F1 score. The traditional networks like YOLO net, seg net, Mask RCNN and U-net obtains less accuracy compared to the Fuzzy u-net. Fuzzy u-net preserves the high accuracy ranges of 97.57%. The proposed MFO-Fuzzy U net model improves the overall accuracy of 3.43%, 0.83% and 9.21% for LinkNet-B7, U-net and FCNs respectively.