vmt This is the implementation of paper Activity recognition with volume motion templates and histograms of 3D gradients [ieee]. It needs lots of refactoring, a proper usage guide and some make-up. I hope to do it when available.