A Registration Method of Cardiac Image

Abstract

The present invention provides a heart image registration method, which includes the following steps: (1) inputting the heart image and determining the initial center position of the heart; (2) determining the heart center position according to the initial center position (3) determine the position of the heart key point according to the initial position of the key point and the position detector of the key point; (4) calculate the correspondence between the key point and the reference image The error sum of the first norm between the key points of the heart, and calculate the error of the norm and the minimum spatial affine transformation, that is, the registration result. The technical solution not only improves the registration speed and accuracy, but also has excellent robustness.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a registration method of heart images. Background technique [0002] At present, for cardiac images, commonly used affine transformation registration methods include manual or semi-automatic methods through human-computer interaction and retrospective registration technology based on voxels. [0003] Manual or semi-automatic methods of human-computer interaction are complicated, inefficient, and there are large differences between observers and observers themselves. [0004] Voxel-based retrospective registration technology is a registration method based on all voxels in the image, which is mainly divided into grayscale-based methods and feature-based methods. Among them, the feature-based method needs to pre-extract some salient features from the atlas image and search for matches in the target image. This approach is often quite challenging. The grayscale-based registration m...

AI Technical Summary

Problems solved by technology

However, the grayscale-based method needs to traverse a large number of voxels for calculation, so this kind of method has a relatively large amount of calculation and low efficiency.

Images