Haskell packages for Computer Vision
This is an experimental Haskell framework for fast prototyping of computer vision and image processing applications. We take advantage of the expressive power of functional programming to write elegant programs supported by standard low level specialized libraries.
Part of this project depended on a noncomercial version of IPP which is no longer available from Intel. We are in the process of replacing this dependency by custom functions, opencv, and IPPICV, an IPP subset freely available for opencv users.
(Tested on Ubuntu 12.04 32bit)
This branch works better with ghc >= 7.8.3
Get the source code:
$ git clone git://github.com/albertoruiz/easyVision.git
Add the following environment variables to
export EASYVISION=/your/path/to/easyVision/ export LD_LIBRARY_PATH=$EASYVISION/lib/lib32
(optional, IPP 7.1) Add the following environment variables to
export IPP_INC=/opt/intel/composerxe/ipp/include export IPP_SHARED="/opt/intel/ipp/lib/ia32/:/opt/intel/lib/ia32/" export IPP_LIBS="ippcore ippi ipps ippcc ippvc ippcv iomp5" export LD_LIBRARY_PATH=$IPP_SHARED:$EASYVISION/lib/lib32 export IPP_LINK=-pthread
Modify as required to make sure that the IPP .h headers are in IPP_INC and the corresponding .so libs are in IPP_SHARED.
Continue installation in a new terminal.
Install the required libraries:
$ sudo apt-get install libgsl0-dev libatlas-base-dev libglpk-dev $ sudo apt-get install mplayer mencoder imagemagick
It is also recommended that you install opencv including the nonfree modules.
Optional: cuda, tesseract, zbar, etc.
Install the Haskell packages:
$ cd easyVision/packages $ make getippicv $ make
The basic system will be correctly installed if several hVision-* packages are shown by
$ ghc-pkg list
Run the demos:
$ cd ../projects/tour $ make play $ ./play
Sorry, at the moment most example programs depend on IPP 7.1.