The wireless capsule endoscope can comprehensively inspect the digestive tract. It has the advantages of safety, painlessness, and no postoperative reaction, but it has some disadvantages. When the most critical issue is, the entire inspection process is very time-consuming. Using image segmentation technology, the digestive tract lesion tissue and surrounding areas can be locally enhanced, and the Region of Interest (ROI) can be segmented, which is helpful for the doctor to read the image. Based on the Mask R-CNN network, this paper proposes a wireless capsule endoscope lesion image segmentation method. Experimental results show that this method can achieve effective segmentation of the capsule lens lesion image.