An open library of computer vision algorithms
Fetching latest commit…
Cannot retrieve the latest commit at this time
|.gitattributes||new Visual C++ project|
|vlfeat.sln||new Visual C++ project|
VisionLab Features Library Andrea Vedaldi and Brian Fulkerson INSTALLING In most cases, you do not need to compile or build the documentation as described below. Instead, to use with MATLAB, simply cd to the toolbox directory and type vl_setup. Binaries are created in the bin/XXX directory, where XXX changes depending on the architecture. VLFeat supports the following (the naming reflects MATLAB naming for MEX files): - maci Mac OS X Intel - maci64 Mac OS X Intel (64 bit) - glx GNU Linux i386 - a64 GNU Linux x64 - w32 Windows - w64 Windows (64 bit) The executable binaries can be installed anywhere in the command line path (the directory contains also the static library). More details may be found in the provided HTML documentation (doc/index.html). COMPILING The core library and command line utilities require a common C compiler supporting the C-90 standard and some C-99 extension (GCC and Visual C will do) plus a few common extension (see APPENDIX). To compile MATLAB support, MATLAB should also be installed and the MATLAB MEX command should be working correctly. In general, issuing > make should be enough to compile VLFeat. Type > make help or refer to http://www.vlfeat.org for further instructions. WINDOWS. VLFeat bundels a Microsoft NMAKE Makefile (Makefile.mak) script that has been tested under Visual C++ 2008 Express and MATLAB R2008a (minor adjustments may be required for other versions). Open the Visual C terminal, cd into the VLFeat directory and issue. > nmake /f Makefile.mak For Windows 64 use > make /f Makefile.mak ARCH=w64 If you do not have Visual C++ 2008, or you have an older version of Matlab and wish to compile the mex files _only_, then you may start from our binary distribution. Enter the toolbox directory: > cd toolbox And run our mex compilation script from the Matlab command line: > vl_compile This has been tested with Matlab R14 and lcc, but other configurations should only require minor tweaking. CREATING THE DOCUMENTATION This task is handled by various Makefile that have been tested under Mac OS X and Linux only. Compiling the documentation requires the following additional tools - fig2dev (part of transfig) - a modern LaTeX with + pdflatex + dvips + htlatex (possibily part of TeX4ht) + dvipng (possibily a separated package) - doxygen FIGURES. Figures are preprocessed by typing > make -C doc/figures However you need to run some MATLAB code to generate part of the figures to start with. To this end, load MATLAB and (provided that everything is compiled and installed correctly) type > vl_demo TUTORIALS. To create the figures for the tutorials, issue > make doc-deep > make doc SOURCE CODE DOCUMENTATION. To compile the source code documentation type > make doc-api See the file doc/index.html for an overview. APPENDIX CODE COMPATIBILITY. In addition to the C-90 standard, the C compiler is supposed to support the following common features: - long int (64 bit integer) support - variadic macro support The SSE accelerated code requires the compiler to support Intel intrisic. GCC and Visual C both satisfy all the requirements. CHANGES 0.9.8 Added image distance transform. Improved documentation. 0.9.7 Changed the Mac OS X binary distribution to require a less recent version of Mac OS X (10.5). 0.9.6 Changed the GNU/Linux binary distribution to require a less recent version of the C library. 0.9.5 Added kd-tree and new SSE-accelerated vector/histogram comparison code. Improved dense SIFT (dsift) implementation. Added Snow Leopard and MATLAB R2009b support. 0.9.4 Added quick shift. Renamed dhog to dsift and improved implementation and documentation. Improved tutorials. Added 64 bit Windows binaries. Many other small changes. 0.9.3 Namespace change (everything begins with a vl_ prefix now). Many other changes to provide compilation support on Windows with Matlab 7. beta-3 Completions to the ikmeans code. beta-2 Many completions. beta-1 Initial public release.