For one of our software projects where investigating the possibilities of the Open Source image processing library_opencv_.
When installing openCV from the ubuntu packages it goes wrong with the python bindings for example. So I followed the instructions at http://opencv.willowgarage.com/wiki/InstallGuide%20%3A%20Debian to get a working version. I used the OpenCV from sourceforge stable. Both the listed SVN version fail to compile at my workstation…
So yes it’s a bitch; but it’s powerfull
Instead of the build instruction with make I had to replace make with cmake:
cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D BUILD_PYTHON_SUPPORT=ON -D BUILD_EXAMPLES=ON ..
Then i did:
make -j4
Then i did a make in the samples/c dir to build all the samples. All the samples build correctly.
This Is the facedetect algorithm; it detects one or more faces in a live video stream.
With some extra work I alos got it working on the MAC; the most easiest way is to download a Universal binary of OpenCV-2.0 with some patches; it will work ass an Mac OSX Framework straight from xCode.
I usedthe description at http://opencv.willowgarage.com/wiki/Mac_OS_X_OpenCV_Port to get it running; I downloaded the DMG for the framework from: http://vislab.cs.vt.edu/~vislab/wiki/images/4/44/OpenCV2.0.dmg
Have fun: this is it working with Snow Leopard:
It will detect all (human) faces with a reasonable accuracy:
