Research on 3D Medical Image Segmentation based on improved 3D-Unet

摘要

Magnetic resonance imaging (MRI) images have high tissue resolution and rich contrast, and are important data for the study of hippocampal morphology. The hippocampal volume can not achieve the ideal segmentation effect in human brain MRI images, and few people have studied the 3D reconstruction of human brain MRI. An improved 3D-Unet medical image processing is proposed in this paper. Through the establishment of 3D segmentation algorithm, the 3D reconstruction and visualization of brain tissue are realized, and the linkage between segmentation algorithm and 3D visualization system is realized. Experiments show that the improved algorithm has high recognition rate and strong adaptability.