3D cranial image registration method based on ROI and conformal geometric algebraic feature invariants

Abstract

The invention discloses a 3D cranial image registration method based on ROI and conformal geometric algebra feature invariants. Cranial region-of-interest contour data serves as a registration point-cloud set, geometric correction is conducted on the x, y and z directions of a reference image and the x, y and z directions of a floating image, and regions of interest of two modalities are captured; effective contour point-cloud sets are adopted to calculate 'minimum projection unit balls' of the point-cloud sets of the two modalities to enable the sum of the projection distances between the point-cloud sets and the unit-ball centers to be minimum, and in this way, a 'first geometric invariant' and the translation amount of the reference image relative to the floating image are determined; then, according to the fact that the sum of the projection distances between the effective contour point-cloud sets and 'minimum unit rings' is the maximum or the minimum, a 'second geometric invariant' and a rotation operator of the floating image relative to the reference image are determined; finally, through the unification of the two geometric invariants, the translation and rotation from the floating image to the reference image are achieved, and a registration effect is reached. The method can be used for accurately positioning three-dimensional locations of tissues and organs, and the execution efficiency and the stability are high.

Description

technical field [0001] The invention relates to a cranial image registration method. Background technique [0002] Existing medical imaging equipment can be divided into anatomical and functional images according to different characteristics of medical images. Anatomical medical maps (such as CT, MRI, and X-ray, etc.) are used to provide information on the anatomical shape of internal organs in the human body, and functional medical maps (such as SPECT and PET, etc.) can reflect the metabolism of internal organ functions. Certain aspects of the human body can be obtained by using corresponding imaging techniques, and different imaging characteristics provide multiple types of diagnostic references. When medical personnel judge the disease, they often need to analyze multiple forms of cross-sectional medical maps of the patient at the same time, so as to obtain the complete anatomical details and functional status of the suspected patient site. In this case, judging multipl...

AI Technical Summary

Problems solved by technology

Existing registration methods such as mutual information, closest point iteration, etc., the huge amount of mutual information data makes the whole registration process take a long time and the calculation is complicated

Images