CVPR
2012
IEEE
11 years 7 months ago
2012
IEEE
We propose an approach for cross-view action recognition by way of ‘virtual views’ that connect the action descriptors extracted from one (source) view to those extracted from...