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 architecture naming for the MEX command the
  MEX file extension differs and is reported in parenthesis):

  - maci     (mexmaci)    Mac OS X Intel
  - maci64   (mexmaci64)  Mac OS X Intel (64 bit)
  - glnx86   (mexglx)     GNU Linux i386
  - glnxa64  (mexa64)     GNU Linux x64
  - win32    (mexw32)     Windows
  - win64    (mexw64)     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

  > nmake /f Makefile.mak ARCH=win64

  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
  0.9.13     Fixes Windows binary package.
  0.9.12     Fixes vl_compile and the architecture string on Linux 32 bit.
  0.9.11     Fixes a compatibility problem on older Mac OS X versions.
             A few bugfixes are included too.
  0.9.10     Improves the homogeneous kernel map. Plenty of small tweaks
             and improvements. Make maci64 the default architecture on
             the Mac.
  0.9.9      Added: sift matching example. Extended Caltech-101
             classification example to use kd-trees.
  0.9.8      Added: image distance transform, PEGASOS, floating point
             K-means, homogeneous kernel maps, a Caltech-101
             classification example. 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.