Learning probabilistic structure for human motion detection
- Creators
- Song, Yang
- Goncalves, Luis
- Perona, Pietro
Abstract
Decomposable triangulated graphs have been shown to be efficient and effective for modeling the probabilistic spatio-temporal structure of brief stretches of human motion. In previous work such model structure was handcrafted by expert human observers and labeled data were needed for parameter learning. We present a method to build automatically the structure of the decomposable triangulated graph from unlabeled data. It is based on maximum-likelihood. Taking the labeling of the data as hidden variables, a variant of the EM algorithm can be applied. A greedy algorithm is developed to search for the optimal structure of the decomposable model based on the (conditional) differential entropy of variables. Our algorithm is demonstrated by learning models of human motion completely automatically from unlabeled real image sequences with clutter and occlusion. Experiments on both motion captured data and grayscale image sequences show that the resulting models perform better than the hand-constructed models.
Additional Information
© 2001 IEEE. Date of Current Version: 15 April 2003. Funded by the NSF Engineering Research Center for Neuromorphic Systems Engineering (CNSE) at Caltech (NSF9402726), and by an NSF National Young Investigator Award to PP (NSF9457618).Attached Files
Published - SONcvpr01.pdf
Files
Name | Size | Download all |
---|---|---|
md5:ac6743f141392cd2eb9a115a01de6d92
|
663.3 kB | Preview Download |
Additional details
- Eprint ID
- 27704
- Resolver ID
- CaltechAUTHORS:20111109-104004085
- EEC-9402726
- NSF
- IIS-9457618
- NSF
- Center for Neuromorphic Systems Engineering, Caltech
- Created
-
2011-11-11Created from EPrint's datestamp field
- Updated
-
2021-11-09Created from EPrint's last_modified field