Abstract | ||
---|---|---|
We present a method that is capable of tracking and es- timating pose of articulated objects in real-time. This is achieved by using a bottom-up approach to detect instances of the object in each frame, these detections are then linked together using a high-level a priori motion model. Unlike other approaches that rely on appearance, our method is entirely dependent on motion; initial low-level part detec- tion is based on how a region moves as opposed to its ap- pearance. This work is best described as Pictorial Struc- tures using motion. A sparse cloud of points extracted us- ing a standard feature tracker are used as observational data, this data contains noise that is not Gaussian in nature but systematic due to tracking errors. Using a probabilistic framework we are able to overcome both corrupt and miss- ing data whilst still inferring new poses from a generative model. Our approach requires no manual initialisation and we show results for a number of complex scenes and differ- ent classes of articulated object, this demonstrates both the robustness and versatility of the presented technique. |
Year | DOI | Venue |
---|---|---|
2008 | 10.1109/CVPR.2008.4587530 | CVPR |
Keywords | Field | DocType |
image motion analysis,object detection,pose estimation,articulated objects,high-level a priori motion model,low-level motion,object detection,pictorial structures,real-time pose estimation | Object detection,Computer vision,Pattern recognition,Computer science,Robustness (computer science),Feature extraction,Pose,Artificial intelligence,Missing data,Articulated body pose estimation,Hidden Markov model,Generative model | Conference |
Volume | Issue | ISSN |
2008 | 1 | 1063-6919 |
Citations | PageRank | References |
6 | 0.44 | 13 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ben Daubney | 1 | 78 | 5.71 |
David Gibson | 2 | 36 | 3.16 |
N.W. Campbell | 3 | 247 | 32.14 |