DMCA
Learning a hierarchy of discriminative space-time neighborhood features for human action recognition (2010)
Cached
Download Links
| Venue: | In IEEE Conference on Computer Vision and Pattern Recognition (CVPR |
| Citations: | 144 - 0 self |
BibTeX
@INPROCEEDINGS{Kovashka10learninga,
author = {Adriana Kovashka and Kristen Grauman},
title = {Learning a hierarchy of discriminative space-time neighborhood features for human action recognition},
booktitle = {In IEEE Conference on Computer Vision and Pattern Recognition (CVPR},
year = {2010}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Recent work shows how to use local spatio-temporal features to learn models of realistic human actions from video. However, existing methods typically rely on a predefined spatial binning of the local descriptors to impose spatial information beyond a pure “bag-of-words ” model, and thus may fail to capture the most informative space-time relationships. We propose to learn the shapes of space-time feature neighborhoods that are most discriminative for a given action category. Given a set of training videos, our method first extracts local motion and appearance features, quantizes them to a visual vocabulary, and then forms candidate neighborhoods consisting of the words associated with nearby points and their orientation with respect to the central interest point. Rather than dictate a particular scaling of the spatial and temporal dimensions to determine which points are near, we show how to learn the class-specific distance functions that form the most informative configurations. Descriptors for these variable-sized neighborhoods are then recursively mapped to higher-level vocabularies, producing a hierarchy of space-time configurations at successively broader scales. Our approach yields state-of-theart performance on the UCF Sports and KTH datasets. 1.
Keyphrases
discriminative space-time neighborhood feature human action recognition informative space-time relationship ucf sport recent work variable-sized neighborhood local motion temporal dimension action category kth datasets particular scaling realistic human action pure bag-of-words model local descriptor training video space-time configuration predefined spatial binning visual vocabulary appearance feature central interest point informative configuration class-specific distance function higher-level vocabulary spatial information approach yield state-of-theart performance nearby point local spatio-temporal feature space-time feature neighborhood
