Skin Segmentation using Deep Learning

Here is a demonstration of real time skin segmentation. In this project a variety of techniques were employed to carry out skin segmentation.  Namely, U-Net, Mask-RCNN, Features and Thresholding.  A detailed study and comparison have been shown in our paper “Adaptation and Evaluation of Deep Learning Techniques for Skin Segmentation on Novel Abdominal Dataset.” 2019 IEEE International Conference on BioInformatics and BioEngineering. IEEE, 2019.

A unique first of its kind abdominal skin dataset was also created to train the neural networks. The dataset consists of 1,400 abdomen images retrieved online from Google images search, which were subsequently manually segmented. The dataset and the code for all the implemented techniques can be found  on Github.