Abstract | ||
---|---|---|
Human pose estimation requires a versatile yet well-constrained spatial model for grouping locally ambiguous parts together to produce a globally consistent hypothesis. Previous works either use local deformable models deviating from a certain template, or use a global mixture representation in the pose space. In this paper, we propose a new hierarchical spatial model that can capture an exponential number of poses with a compact mixture representation on each part. Using latent nodes, it can represent high-order spatial relationship among parts with exact inference. Different from recent hierarchical models that associate each latent node to a mixture of appearance templates (like HoG), we use the hierarchical structure as a pure spatial prior avoiding the large and often confounding appearance space. We verify the effectiveness of this model in three ways. First, samples representing human-like poses can be drawn from our model, showing its ability to capture high-order dependencies of parts. Second, our model achieves accurate reconstruction of unseen poses compared to a nearest neighbor pose representation. Finally, our model achieves state-of-art performance on three challenging datasets, and substantially outperforms recent hierarchical models. |
Year | DOI | Venue |
---|---|---|
2012 | 10.1007/978-3-642-33715-4_19 | ECCV (5) |
Keywords | Field | DocType |
compact mixture representation,recent hierarchical model,spatial hierarchy,latent node,local deformable model,global mixture representation,hierarchical structure,well-constrained spatial model,high-order spatial relationship,new hierarchical spatial model,mixture model,pure spatial | Computer science,Pose,Artificial intelligence,Hierarchy,Hierarchical database model,k-nearest neighbors algorithm,Computer vision,Exponential function,Pattern recognition,Inference,Tree (data structure),Mixture model,Machine learning | Conference |
Volume | ISSN | Citations |
7576 | 0302-9743 | 62 |
PageRank | References | Authors |
2.80 | 15 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yuandong Tian | 1 | 703 | 43.06 |
C. Lawrence Zitnick | 2 | 7321 | 332.72 |
Narasimhan, S.G. | 3 | 2348 | 169.35 |