Motion History Volumes and Action Recognition from Arbitrary Views using 3D Exemplars

- "Motion History Volumes" (MHV) are a free-viewpoint representation for human actions in the case of multiple calibrated, and background-subtracted, video cameras. I will present algorithms based on MHVs for segmentation and classification of different actions performed by different people in a variety of viewpoints.
- "Action Recognition from Arbitrary Views using 3D Exemplars" In this work we address the problem of recognizing actions from arbitrary views, even from a single camera. We propose a new framework where we model actions using three dimensional occupancy grids, built from multiple viewpoints, in an exemplar-based HMM. The novelty is, that a 3D reconstruction is not required during the recognition phase, instead learned 3D exemplars are used to produce 2D image information that is compared to the observations.