Software for ABSORB: Atlas Building by Self-Organized Registration and Bundling


0. Download packages


Software package for download

To help us improve the software, we need your feedback. Please send the following information to jiahj@med.unc.edu, and all information collected will be kept confidential and used for the purpose of improving the software. Please cite the paper of ABSORB [1] as in References if your publishing studies use ABSORB software.

       Your Name / Email / Institution / Purpose of Download / From Where You Hear Us

1. Purpose


This software package implements ABSORB: Atlas Building by Self-Organized Registration and Bundling [1,2] – an algorithm for effective groupwise registration, which has been published as:

Hongjun Jia, Guorong Wu, Qian Wang, Dinggang Shen, "ABSORB: Atlas Building by Self-Organized Registration and Bundling", NeuroImage, Vol. 51, No. 3, 1 July 2010, pp. 1057-1070. [PDF file]

The package is available free to the public for the academic research purpose at http://bric.unc.edu/ideagroup/free-softwares/. This software is distributed WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

 

ABSORB_framework
2. System requirements


This software is developed using ITK and has been tested on Windows XP (32-bit) and Linux (64-bit, kernel version 2.6.18-194.el5).

The required input is a set of 3D MR intensity images (in Analyze format with paired .hdr and .img files) with a text file (.txt) listing all header file (.hdr) names. The output is the set of registered images together with the corresponding dense deformation fields.

In ABSORB, we used the following software packages, affinity propagation, diffeomorphic demons, and tcl/tk. Please refer to the software manual for details.

 

3. Image pre-processing

The images should be pre-processed before applying ABSORB:

All brain MR images used as inputs to ABSORB should be in the same situation (e.g., skull-stripped or not, cerebellum removed or not, etc.).

The input images should be in Analyze format with paired header and image files.


4. Software package

(download links are here.)

Package for win-32bit includes:

ABSORB_GUI_main.exe, ABSORB_command.exe, apclusterwin.dll, ApplyDeformationField.exe, MSVCP80.dll, MSVCR80.dll, MSVCM80.dll,

tcl85.dll, tk85.dll, a tcl/tk library folder: lib, and vcredist_x86.exe

 

Package for linux-64bit includes:

ABSORB_GUI_main, ABSORB_command, apclusterunix64.so and ApplyDeformationField


Test data for ABSORB includes:

8 MR brain images in Analyze format (128*128*128) with header file list, and they have NOT been affinely registered.


5. Software document


Please read lastest ABSORB software manual.


6. References


[1] Hongjun Jia, Guorong Wu, Qian Wang, Dinggang Shen, "ABSORB: Atlas Building by Self-Organized Registration and Bundling", NeuroImage, Vol. 51, No. 3, 1 July 2010, pp. 1057-1070.

[2] Hongjun Jia, Guorong Wu, Qian Wang, Dinggang Shen, "ABSORB: Atlas Building by Self-Organized Registration and Bundling", in Proc. of the 23rd IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2010), San Francisco, CA, 2010.


7. Contact us


For any questions or bug reports, please email jiahj@med.unc.edu.

 

8. Update history

 

20100901   Update manual

20100729   Minor update

20100722   Add options for Affine Registration and Output data type

20100714   Update GUI

20100712   Update document

 

This page last updated on Oct. 29, 2010.

Filed under: