Important note: these binaries can be used for non-commercial purposes only. Purpose: Scale Invariant Interest Point Detectors, Local Descriptors They are used in the paper: Gy. Dorkó and C. Schmid Selection of Scale Invariant Neighborhoods for Object Class Recognition 9th International Conference on Computer Vision , 2003. (http://www.inrialpes.fr/lear/people/dorko/pub.html) This archive contains the following binaries: Detect - Scale invariant interest point detection ComputeDescriptor - computing local descriptors DrawCorners - Show the result on the image dumpcontents - convert the result to ascii text hldesc - print the descriptors (obsolete in favor of corners2text) corners2matlab-r12 and corners2matlab-r13 - convert the result to matlab format corners2text - convert detected points/descriptors to text MergeCorners - merge detected files SelectCorners - select some descriptors from the binary files text2corners - convert text files to the input format of ComputeDescriptor ipld.changelog.txt - changes to the archive ipld.readme.txt - this readme file General binary information: Most of them: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.2.5, statically linked Shortly: If you have a pc with Linux installed they should all work. corners2matlab-xx ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.2.5, dynamically linked (uses shared libs), not stripped Dynamically linked libraries: linux-gate.so.1 libmx.so libut.so libmat.so libmi.so libstdc++.so.6 libm.so.6 libgcc_s.so.1 libc.so.6 libstdc++-libc6.1-2.so.3 libdl.so.2 libpthread.so.0 /lib/ld-linux.so.2 You must have all of these to run the matlab conversation tool. (Don't forget to set you LD_LIBRARY_PATH variable if it is necessary.) Shortly: You should not have any problem if gcc 3.x and MatLab installed on your system. You may need to get libstdc++-libc6.1-2.so.3 from somewhere (matlab libs use it). Note that there are 2 matlab conversion utility corners2matlab-r12 and corners2matlab-r13 If you have a different version try to use corners2text and load the text file in matlab. For general help and available options please run the appropriate binary with -help option.