Registration image determination method and registration image determination device

Abstract

The invention provides a registration image determination method and a registration image determination device, and the method comprises the steps: determining a speed field function set correspondingto a to-be-registered pixel unit set in a to-be-registered image based on a target image and the to-be-registered image, wherein the to-be-registered pixel unit set comprises a plurality of to-be-registered pixel units, the speed field function set comprises a plurality of speed field functions, and the plurality of to-be-registered pixel units and the plurality of speed field functions have a one-to-one correspondence relationship; and determining a registration image corresponding to the to-be-registered image based on the speed field function set and the target image. According to the technical scheme, a flow registration method in a conventional algorithm is optimized, regularization is carried out on the speed, and then the purpose of large-deformation nonlinear registration is achieved on the premise that the registration speed is guaranteed.

Description

technical field [0001] The present application relates to the technical field of medical image registration, and in particular to a method for determining a registered image and a device for determining a registered image. Background technique [0002] As one of the traditional algorithm registration methods, the flow registration method can deal with the nonlinear registration problem with large deformation, but the processing speed of the flow registration method is slow, and the features in the processing process are artificially specified energy features. Far inferior to features automatically extracted in deep learning algorithms. The elastic registration in the deep learning method registration is limited by the regular term in the displacement field, which also makes elastic registration unable to solve the nonlinear registration problem with large deformation. [0003] Based on this, how to not only solve the nonlinear registration problem with large deformation, bu...

AI Technical Summary

Problems solved by technology

[0002] As one of the traditional algorithm registration methods, the flow registration method can deal with the nonlinear registration problem with large deformation, but the processing speed of the flow registration method is slow, and the features in the processing process are artificially specified energy features. Far inferior to the features automatically extracted in deep learning algorithms

However, the elastic registration in the deep learning method registration is limited by the regular term in the displacement field, which also makes elastic registration unable to solve the nonlinear registration problem of large deformation.

Images