What is this?
zxhproj (doxygen html) is a medical image computing platform, being developed and maintained by Xiahai Zhuang since 2004. Based on it, several image registration and segmentation tools have been developed, such as multivariate mixture model, cardiac segmentation, locally affine registration method (LARM), spatially encoded mutual information (SEMI), as well as other image/vector field processing tools. The tools work with nifty (nii/nii.gz) and gipl short int data formats.

The tools released are for research purpose ONLY and not free from bugs, thus should NOT be used in any clinical related situations. Any usage is entirely at your own risk.

Download all

Below are helps.
The binary (executable) tools without argument or with argument -H (in detail) will give the usage of the tool.
zxhmvmmseg (paper: T PAMI): multivariate mixture model for simultaneous registration and segmentation combining multi-source images; the cplusplus source code in github.
atlas-based segmentation
multi-modality multi-atlas segmentation and label fusion
zxhLabelFuse (papers: MedIA MP): this tool provides the multi-scale label fusion methods for multi-modality multi-atlas segmentation
Usage: zxhLabelFuse TargetImage.nii TargetLabelSavePre nAtlas ImageAtlasList.txt LabelAtlasList.txt -method1
zxhregsemi0 (paper): nonrigid registration - spatially encoded mutual information (SEMI) using free-form deformations, e.g.:
zxhregsemi0 -target target.nii -source source.nii -o output -Reg 3 -steps 50 50 20 -sub 4 4 4 -sub 4 4 4 -sub 2 2 2 -ffd 40 40 40 -ffd 20 20 20 -ffd 10 10 10 -bending 0.001
OR diffeomorphism without bending energy -bending 0
Nonrigid registration using other similarities, cross correlation (-cc), sum square difference (-ssd), e.g.:
zxhregsemi0 -test target.nii -ref source.nii -o output -Reg 3 -sub 4 4 4 -sub 4 4 4 -sub 2 2 2 -ffd 40 40 40 -ffd 20 20 20 -ffd 10 10 10 -bending 0.01 0.005 0.001 -cc
zxhreglarm/zxhclasslarm (paper): locally affine registration method, LARM, e.g.:
zxhreglarm -test target.nii -ref source.nii -o output -localmatrix 3 -localregion localregion1.nii localregion2.nii localregion3.nii -Reg 5 -sub 4 4 4 -sub 4 4 4 -sub 3 3 3 -sub 3 3 3 -sub 2 2 2 -length 1 1 0.5 0.5 0.3
For cardiac, use option -heart and resample images into 2x2x2 mm using zxhtransform (e.g.: zxhtransform originalimage.nii -o saveimage.nii -resave -spacing 2 2 2 -bspline)
zxhdraw (paper): invert displacement using Dynamic Resampling And distance Weighting interpolation, DRAW, e.g.:
zxhdraw refimage.nii -o saveinversedisplace -n 1 -d displace_x.nii displace_y.nii displace_z.nii -inv 1
zxhreg: rigid, similarity (scale), affine registration, e.g.:
zxhreg -test target.nii -ref source.nii -o output -pre 12
-pre 12: pre-align center of two images;
-pre 13: pre-align gray level center of two images;
zxhcmrsc (paper): cardiac MRI, cine short-axis and long-axis shift correction
zxhtransform: transform source image to target space using registration results/transformations
zxhtransformvector: compute transformation fields into some thing
zxhimageop: operations on images
zxhimageinfo: image info, convert to float/short int image

zxhproj, version 2.2 (c) 2004-2017 ZHUANG, XiaHai (zhuangxiahai@163.com; xiahai.zhuang@gmail.com). All rights reserved.