DocumentCode :
595446
Title :
Robust 3D human pose estimation via dual dictionaries learning
Author :
Hao Ji ; Fei Su
Author_Institution :
Sch. of Inf. & Commun. Eng., Beijing Univ. of Posts & Telecommun., Beijing, China
fYear :
2012
fDate :
11-15 Nov. 2012
Firstpage :
3370
Lastpage :
3373
Abstract :
In this paper, a new dual dictionaries learning (DDL) method is proposed for robust 3D human pose estimation. The performance and applicability of traditional methods are limited by a lack of robustness to corrupted observations caused by occlusions or poor background subtraction. Our DDL approach aims at simultaneously constructing two overcomplete dictionaries, called the visual observation dictionary (VBD) and the body configuration dictionary (BCD), with a shared sparse representation (SSR) regularization with respect to every data sample. Under such regularization, two dictionaries are tied together and the 3D pose estimation problem can be reduced to a simple ℓ1 optimization problem given a new test visual observation. We also propose a efficient algorithm based on inexact Augmented Lagrange Multiplier (IALM) method to solve the above DDL optimization model. Experimental results on HumanEva database show the superiority of our approach over several current state of the art methods.
Keywords :
dictionaries; hidden feature removal; image representation; learning (artificial intelligence); optimisation; pose estimation; visual databases; BCD; DDL optimization model; HumanEva database; IALM method; SSR regularization; VBD; augmented Lagrange multiplier method; body configuration dictionary; data sample; dual dictionaries learning; l1 optimization problem; occlusions; overcomplete dictionaries; poor background subtraction; robust 3D human pose estimation; shared sparse representation regularization; test visual observation; visual observation dictionary; Dictionaries; Estimation; Humans; Optimization; Robustness; Sparse matrices; Visualization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pattern Recognition (ICPR), 2012 21st International Conference on
Conference_Location :
Tsukuba
ISSN :
1051-4651
Print_ISBN :
978-1-4673-2216-4
Type :
conf
Filename :
6460887
Link To Document :
بازگشت