Wiki

Clone wiki

openrecon / Home

Welcome to OpenRecon !

OpenRecon is a 2D image reconstruction toolbox for Matlab. Written by C++ the initial motivation for this project is to make CT reconstruction easy and fast to implement for scientific research and development. One is able to test a CT problem from projection simulation to reconstruction by utilizing simple Matlab commands (eg. first typing

sino=xprj(img);

to generate a sinogram, and then typing

rec_img=xrec(sino);

to issue a reconstruction). Various parameters can be appended while launching these commands. Currently it integrates projection simulations and reconstruction algorithms for different scanning modes. CT problems include parallel-beam/fan-beam scanning geometries, limited-angle scan, few-view scan, interior problem, bad detector beams, random angle scan, and so on. Algorithms include direct back-projection, filtered back-projection (FBP), SART, OS-SART, SIRT, EM, OS-EM. Besides the toolbox supports different Compressed Sensing minimization techniques including total variation (TV), soft-thresholding (STH) and anisotropic total variation (ATV).

The fork from openct project is mainly for name-change to avoid conflict with existing open-source library of same name. Also all new developments are shifted to this fork. Note in current available branches [default] is maintained by Kriti Sen Sharma with PDF manuals and [xjin local] is a branch that was being maintained by Xin Jin on a local repository. The two versions are being merged gradually. To get one copy of either version, click [Downloads] tab on top of the page. Then in the [Branch] section, choose a branch and select your familiar archive format (zip, gz, etc.) on the right side.

Before using the toolbox it has to be built with Matlab command mex. Please first run

mex -setup

on Matlab command window to make configuration. You may easily get more information on this from google. After that run script build.m which is contained in the archive to compile the toolbox. It is suggested that you install VS2008 under windows OS and install gcc 4.2 under Linux OS.

Contributors: [Kriti Sen Sharma and Xin Jin]

Updated