A multi-atlas fusion method, system and medical equipment

Abstract

The invention provides a method, a system and a medical device of multiple atlas fusion. The method, the system and the medical device of the multiple atlas fusion are suitable for a multiple atlas system. The method includes that registration of multiple groups of atlas images and a target image is carried out, the registered multiple atlas images and target image are divided into sub-block regions, local correlation between each sub-block region of the target gray level image and the corresponding sub-block region in each atlas gray level image is calculated, and in the corresponding local correlation result of each sub-block region, the atlas gray level images which are located top X in terms of correlation values are selected, wherein the X is a positive integer. According to atlas identification images which correspond to the atlas gray level images which are located top X, each sub-block region of the target gray level image is fused. The fusion result of each sub-block region is integrated so that a fusion image of the target image is obtained. According to the method, the system and the medical device of the multiple atlas fusion, atlas local information with a good registration effect is fully considered, and therefore a segmentation effect can be achieved more accurately and quickly.

Description

technical field [0001] The invention belongs to the technical field of medical image processing, and in particular relates to a multi-atlas fusion method, system and medical equipment. Background technique [0002] The most important purpose of map fusion is to provide an effective, reliable and accurate method for a series of divided maps to be fused into the resulting image. Multi-map fusion is an important method of image segmentation. [0003] With the continuous improvement of computer processing power, map fusion technology continues to develop, and map fusion methods are also constantly updated. For example, according to the current popular AdaBoost algorithm, a series of "weak classifiers" are combined into a "strong classifier". A fusion algorithm that uses classifiers to classify maps is proposed; based on the EM algorithm in machine learning, a STAPLE fusion algorithm is proposed, which can simultaneously estimate the performance of the input segmentation map and ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the embodiments of the present invention is to provide a multi-atlas fusion method, which aims to solve the problem that most of the existing atlas fusion algorithms are based on the overall global registration image, and the images with better local registration are not fully considered, which makes the local registration The problem that the map with good effect does not play a role

Images