Training Data Generation for U-Net Based MRI Image Segmentation using Level-Set Methods

Image segmentation has been a well-addressed problem in pattern recognition for the last few decades. As sub-problem of image segmentation, background separation biomedical images generated by magnetic resonance imaging (MRI) also interest applied mathematics literature. Level set evolution active contours idea can successfully be to MRI extract region (ROI) as crucial preprocessing step medical analysis. In this study, we use classical level solution create binary masks various brain which black color implies and white ROI. We further used mask pairs train deep neural network (DNN) architecture called U-Net, proven successful model segmentation. Our experiments have shown that properly trained U-Net achieve matching performance method. Hence were able using automatically input label data successfully. The detect ROI faster than level-set method tool more enhanced analysis studies.