From: Abhishek D. <the...@gm...> - 2008-12-20 14:55:09
|
Hi all, After lots of searching and testing I finally made choice to use the "vigra<http://kogs-www.informatik.uni-hamburg.de/%7Ekoethe/vigra/>" library (http://kogs-www.informatik.uni-hamburg.de/~koethe/vigra/). Despite its weird name, it has everything that a computer vision application would every want. The most interesting part is that everything is Generic which means that eveything is easy to glue together. There are very less tutorials and the only source of information is example programs, codes in the unit test and the doxygen documentation. It was a bit difficult to get started but everything is a breeze once you get used to it. After considering the OpenCV, Octave and GSL (GNU Scientific Library) I decided to use this. GSL is a good choice, but after using it I found that it is designed with focus on mathematical stuffs rather than things that will be mostly used by computer vision applications. Anjan and Bibek must be gearing up to start the C++ implementation of Adaboost. I am sure you will also be searching for a library that helps perform computer vision related tasks. As we have already implemented the algorithm in Matlab, we don't need to bother to write boilerplate codes to do some basic things. So using a library is a good choice. You can give "vigra" a try. I you need any help I can write a short HOWTO in our WIKI. bye -- Abhishek http://adutta.np.googlepages.com |