The Community for Technology Leaders
2012 IEEE Conference on Computer Vision and Pattern Recognition (2012)
Providence, RI USA
June 16, 2012 to June 21, 2012
ISSN: 1063-6919
ISBN: 978-1-4673-1226-4
pp: 1218-1225
L. S. Davis , Univ. of Maryland, College Park, MD, USA
V. I. Morariu , Univ. of Maryland, College Park, MD, USA
S. Khamis , Univ. of Maryland, College Park, MD, USA
ABSTRACT
We propose a framework that performs action recognition and identity maintenance of multiple targets simultaneously. Instead of first establishing tracks using an appearance model and then performing action recognition, we construct a network flow-based model that links detected bounding boxes across video frames while inferring activities, thus integrating identity maintenance and action recognition. Inference in our model reduces to a constrained minimum cost flow problem, which we solve exactly and efficiently. By leveraging both appearance similarity and action transition likelihoods, our model improves on state-of-the-art results on action recognition for two datasets.
INDEX TERMS
gesture recognition, action transition likelihood, flow model, joint action recognition, identity maintenance, multiple targets, appearance model, network flow-based model, bounding boxes, video frames, constrained minimum cost flow problem, Humans, Mathematical model, Training, Context, Maintenance engineering, Equations, Joints
CITATION

L. S. Davis, V. I. Morariu and S. Khamis, "A flow model for joint action recognition and identity maintenance," 2012 IEEE Conference on Computer Vision and Pattern Recognition(CVPR), Providence, RI USA, 2012, pp. 1218-1225.
doi:10.1109/CVPR.2012.6247804
232 ms
(Ver 3.3 (11022016))