Research on 3D Contrast Image Segmentation Based on an Improved Otsu Algorithm

Abstract

Abstract This article mainly introduces typical segmentation methods of medical images, especially when using the advantages of threshold segmentation when segmenting blood vessels, and it is effective when selecting the principle of maximum between-class variance and global threshold. Sex is explained. Based on this, the key role played by the local threshold in the segmentation of angiographic images is explained. In addition, the block method used in this article is a fixed size. Considering when the block size can show the best segmentation effect, the article proposes an index that can determine the block size, namely block the rate of change of the internal standard deviation of the image, and the proposed method is also demonstrated.