Abstract | ||
---|---|---|
This paper proposes a generative method to extract 3D human pose using just a single image. Unlike many existing approaches we assume that accurate foreground background segmentation is not possible and do not use binary silhouettes. A stochastic method is used to search the pose space and the posterior distribution is maximized using Expectation Maximization (EM). It is assumed that some knowledge is known a priori about the position, scale and orientation of the person present and we specifically develop an approach to exploit this. The result is that we can learn a more constrained prior without having to sacrifice its generality to a specific action type. A single prior is learnt using all actions in the Human Eva dataset [9] and we provide quantitative results for images selected across all action categories and subjects, captured from differing viewpoints. |
Year | DOI | Venue |
---|---|---|
2010 | 10.1109/ICPR.2010.840 | ICPR |
Keywords | Field | DocType |
accurate foreground background segmentation,stochastic method,human eva dataset,specific action type,existing approach,binary silhouette,generative method,human pose,single image,action category,iterative refinement,expectation maximization,pose estimation,stochastic processes,image segmentation,feature extraction,foreground background,posterior distribution | Iterative refinement,Computer vision,Pattern recognition,Segmentation,Expectation–maximization algorithm,Computer science,Image segmentation,Feature extraction,Posterior probability,Pose,Artificial intelligence,Foreground-background | Conference |
Citations | PageRank | References |
1 | 0.36 | 6 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ben Daubney | 1 | 78 | 5.71 |
Xianghua Xie | 2 | 383 | 37.13 |