A robust similarity measure for volumetric image registration with outliers

作者：

Highlights：

• Investigation of robust alignment methods for volumetric medical images

• Proposal of two novel similarity measures based on the cosine of normalised 3D volumetric gradients

• The measures are shown to be robust to occlusions and bias field corruption.

• The first review of robust 3D Lucas–Kanade methods for affine image alignment

• The proposed methods show good performance for non-rigid alignment.

摘要

•Investigation of robust alignment methods for volumetric medical images•Proposal of two novel similarity measures based on the cosine of normalised 3D volumetric gradients•The measures are shown to be robust to occlusions and bias field corruption.•The first review of robust 3D Lucas–Kanade methods for affine image alignment•The proposed methods show good performance for non-rigid alignment.

论文关键词：Image registration,Lucas–Kanade,Normalised gradient,Free-form deformation

论文评审过程：Received 9 February 2015, Revised 30 April 2016, Accepted 5 May 2016, Available online 29 May 2016, Version of Record 12 June 2016.

论文官网地址：https://doi.org/10.1016/j.imavis.2016.05.006