MR image segmentation and displacement field correction method based on geodesic models of local and global areas

Abstract

The invention belongs to the field of MR image segmentation and application, and relates to an MR image segmentation and displacement field correction method based on geodesic models of local and global areas. Based on the critical importance of MR image segmentation in medical image analysis, even more difficult segmentation caused by MR image grayscale non-uniformity and grayscale non-uniformity caused by noise and displacement field, global and local information of images extracted based on global and local symbol pressure functions are used to process images of grayscale non-uniformity; a displacement field correction item is added to the local symbol pressure function to perform MR image segmentation and displacement field correction simultaneously so as to overcome grayscale non-uniformity caused by the displacement field; and the models are expanded from a two-item level set to a four-item level set to realize accurate segmentation of gray matter, white matter and cerebrospinal fluid of a brain MR image. The method is applied to synthetic images and MR images, and the segmentation results show that the method is of significant accuracy and efficiency.

Description

technical field [0001] The invention belongs to the field of image segmentation (Magnetic Resonance, MR) and its application. It involves new level-set-based segmentation algorithms, and in particular a method based on geodesic models of local and global regions that enables simultaneous MR image segmentation and offset field correction. Background technique [0002] The prior art discloses that the level set method (levelset method) is a method for solving curve evolution, which expresses planar closed curves or three-dimensional closed surfaces in an implicit way, thereby avoiding tracking during the evolution process of closed curves, and will Curve evolution is transformed into a purely numerical problem of solving partial differential equations. In recent years, the level set method has gradually become a research hotspot, and has been applied to image segmentation, image smoothing, motion segmentation, that is, moving object tracking, and even stereo vision and image ...

AI Technical Summary

Problems solved by technology

So far, there are many MR image segmentation methods, but the segmentation accuracy is still not very high

Images