Priori-knowledge-based reconstruction method of closed skeleton contour in medical fault image

Abstract

Disclosed in the invention is a priori-knowledge-based reconstruction method of a closed skeleton contour in a medical fault image. Pretreatment is carried out on skeleton contours extracted from a medical fault image, and a single-loop skeleton contour with no redundant branch lines and intersection situation is used as a reference template and others are used as non-reference templates; with introduction of a same-type skeleton template concept, a contour matching the reference template best is searched from the same-type skeleton template by using a characteristic value method and affine transformation is established, and same-type skeleton templates matching the non-reference templates are found out at the same direction by using the neighboring relation between the non-reference templates and the reference template; and then the reference template and the same-type skeleton templates are used as priori knowledge to complete broken edges of the non-reference templates and removing uncorrelated contour lines. According to the invention, frequently occurring problems of intersection, neighbouring, and edge breaking and the like during contour extraction are classified and discussed, thereby substantially improving the quality and efficiency of contour extraction and generating a significance in improvement of precision of three-dimensional reconstruction.

Description

technical field

[0001] The invention belongs to the technical field of computer graphics, and in particular relates to bone outlines to be extracted during the three-dimensional reconstruction of bones in medical tomographic images. Background technique

[0002] In the process of three-dimensional bone reconstruction of tomographic images, bone contour recognition is particularly important, which is directly related to the accuracy of the final reconstruction results. It cannot truly restore its original appearance. There are many literatures at home and abroad dedicated to finding better methods to identify image contours, such as using adaptive methods to automatically obtain high and low thresholds in the edge detection stage, and combining multiple detection techniques to obtain the best edge contours; In the stage, the mean value of the gray values ​​on both sides of the edge point is used as an edge contrast feature to identify the direction of the edge line, or the a...

Images