An implementation of SVM+ with MATLAB QP solver. It has also been tested with MOSEK QP solver.
A faster implementation of kernel SVM+ based on a new SVM+ formulation is also released. You need libsvm to run it.
If you feel it is useful, please cite the following papers:
Wen Li, Dengxin Dai, Mingkui Tan, Dong Xu, and Luc Van Gool, “Fast Algorithms for Linear and Kernel SVM+,” IEEE International Conference on Computer Vision and Pattern Recognition(CVPR),2016
For any question, please contact Wen Li via liwenbnu@gmail.com.
Dependencies
The libsvm library is needed. I have included a compiled mex file (Windows64 version). For other platform, please
- download the latest libsvm package, and run "<LIBSVM_ROOT>/matlab/make.m" to comiple the mex file compatiable to your OS.
- Put the obtained mex file in the folder of libsvm+, or add the folder containing mex file to your matlab paths at the beginning of demo_mnist_svmplus.m
addpath('<LIBSVM_ROOT>/matlab/')
How to use
Simple. Run "demo_mnist_svmplus.m", and see the results^_^.
Copyright
Non-commercial use only. All rights reserved.