I am currently working on a project that requires me to segment the skin layer out of MRI-scans, clean up the image and write this to a surface file for further processing. However, I haven’t got much experience with 3D medical image processing (I’ve been dabbling along the surface of VTK for the past year) and hardly any experience with image processing altogether. As the title suggests, the scans are supplied in the Analyze format (.hdr header and .img data), while the segmentation algorithm I’m running on them needs MetaImage (.mhd header and .raw data) input files. At the moment, I’m converting each scan manually by opening it in Volview and saving is as a mhd volume. As you can image, this is becoming increasingly tedious, so I am looking for the best way to automate the process.
I using Windows 7 64-bit and I have working versions of VTK 5.10.0.RC1, MinGw and Python 2.7 installed (and Matlab R2011a, though I’m not currently using it). Google shows possible solutions using ITK and Matlab. However, it has been such a hassle to install the above software packages and get them to work together that I’m extremely reluctant to change anything to the current setup. My planning doesn’t alow the time to be stuck debugging cryptic compilation errors for the next few weeks.
My question is: what software or method will get me to automate the conversion of Analyze to MetaImage files as fast and painless (installation-wise) as possible?
I can really recommend the DeVIDE software package for things like this (http://code.google.com/p/devide/). It includes everything you need, including a Python distribution that includes cmake, swig, Python, numpy, matplotlib, wxPython, gdcm, VTK, ITK and DeVIDE itself. Furthermore, you can rapidly test and prototype your solution via visual programming.
I’m not entirely sure, but I do believe you need ITK for the Analyze format and if you don’t want to go through the trouble of installing, it is included in DeVIDE as well.