Segm 0.90 reviewDownload
Segm program for segmentation of MR images of the brain. Segm is a set of image analysis tools developed at and used by the Brain
Segm program for segmentation of MR images of the brain.
Segm is a set of image analysis tools developed at and used by the Brain Behavior Laboratory of the University of Pennsylvania. It consists of a set of programs (BrainMask, Kmean_3Dseg, and AdpKmean_3Dseg_Ebeta) originally developed by Michelle Yan for use with specific MR imaging protocols used at the Brain Behavior Lab.
They are intended for the quantitative analysis of MRI volume images of the human brain, and particularly for automatic segmentation into tissue types. While these programs have shown themselves very effective for the particular MR imaging protocols they were designed for, they have not proven adaptable to other imaging protocols. Thus, while they continue to be heavily used at BBL for image analysis, they are not recommended for new applications. We do not anticipate making any further changes to segm, except perhaps to fix crippling bugs or to support data interchange with newer and more portable programs.
BBLimage is BBL's newer toolkit for MR image and may be of interest to users of segm; it is, in any case, often preferable to segm for new applications.
Image files are always read and written in big-endian order, regardless of the native architecture; this was done because most or all of the image files at BBL have been generated and stored on the Sparc workstation, and it seemed easier to have programs on all other
platforms read and write this format than to be constantly converting data files.
Segm has been sucessfully compiled and run under Red Hat Linux 6.2 for Intel and Solaris 2.7 for Sparc. (The current version requires the gcc compiler, mostly because we haven't had the time to figure out exactly how to compile shared library under Sun's cc compiler.) We will be interested to hear about other successes or failures, and very interested to receive fixes that will yield success on other machines and operating systems.
If you are installing both BBLimage and segm, you should install segm first, then BBLimage; there are a few installed files that are shared between these two packages, and BBLimage is most likely to have the most recent version.
The following steps will usually suffice:
1. Unpack the tar file and cd into the source tree.
2. Run "./configure" to guess the right parameters for your machine.
By default, `make install' will install the package's files in `/usr/local/bin', `/usr/local/man', and `/usr/local/lib'. You can specify an installation prefix other than `/usr/local' by giving `configure' the option `--prefix=PATH'. For example, use
to install the programs in /home/hughett/bin, the man pages in /home/hughett/man, and the shared library in /home/hughett/lib/
3. Run "make" to compile and link everything.
4. If you're a programmer, run "make tags" to create the TAGS file for emacs; if you're a vi partisan, make the obvious change to the Makefile first.
5. As root, run "make install" to install all the executable binaries and man pages, typically in /usr/local/bin and /usr/local/man. (The macro PREFIX in the Makefile controls where the files are installed; you might want to look at this to insure that configure has guessed correctly.)
6. You may also need to modify the search paths for binaries and man pages (usually PATH and MANPATH).
7. The shared library libsegm.so is installed in $PREFIX/lib; you may need to add this directory to the search path (usually LD_LIBRARY_PATH) for shared libraries.
Segm 0.90 search tags