CT Image Segmentation of Rectal Tumor Based on U-Net Network

摘要

With the development of artificial intelligence technology, deep learning is more and more widely used in medical image segmentation, and has achieved better segmentation results than traditional segmentation techniques. In this paper, the computer tomography (CT) images of rectal tumors are automatically segmented based on U-Net network. The Jaccard index of the training set is 0.9703, and the Jaccard index of the validation set is 0.9847. Experiments show that U-Net has better effect on the segmentation of rectal tumor images, and it also proves that U-Net has better effect on the segmentation of medical images in small data sets.