Boost CVPR
Boost based
Computer Vision and Pattern Recognition Library
Boostcvpr is a collection of C++
classes, functions and templates that implement many state-of-the-art popular
algorithms of Image Processing, Pattern Recognition and Computer Vision. It is
based on the famous Boost C++ library (http://www.boost.org/).
The key feature that distinguished Boostcvpr from other related libraries like
OpenCV is that it is a pure C++ library and takes full advantages of C++
templates and Boost generic programming. It does not rely on external numerical
libraries.
Boostcvpr uses boost::multi_array<> to contain images and matrices,
boost::graph <> to represent graph, and boost::lambda and boost::mpl to build
compile-time algorithms. This will make Boostcvpr very flexible and versatile.
The algorithms in Boostcvpr can be roughly divided into four categories:
*Basic algorithms, such as eigen solver, graph manipulation, fitting and
interpolation.
*Image processing, such as filtering, histogram, edge detection.
*Pattern recognition and statistical learning, such as PCA, ICA, LDA,
meanshift, SVM.
*Computer vision such as optical flow, kalman filter, image segmentation,
object tracking.
The audience of Boostcvpr is researchers, students and engineers working in
artificial intelligence, machine learning and computer vision.