Abstract | ||
---|---|---|
We describe a probabilistic framework for detecting and tracking motion boundaries. It builds on previous work [4] that used a particle filter to compute a posterior distribution over multiple, local motion models, one of which was specific for motion boundaries. We extend that framework in two ways: 1) with an enhanced likelihood that combines motion and edge support, 2) with a spatiotemporal model that propagates beliefs between adjoining image neighborhoods to encourage boundary continuity and provide better temporal predictions for motion boundaries. Approximate inference is achieved with a combination of tools: Sampled representations allow us to represent multimodal non-Gaussian distributions and to apply nonlinear dynamics. Mixture models are used to simplify the computation of joint prediction distributions. |
Year | DOI | Venue |
---|---|---|
2001 | 10.1109/CVPR.2001.990983 | 2001 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, VOL 2, PROCEEDINGS |
Keywords | Field | DocType |
motion estimation,predictive distribution,nonlinear optics,probability,mixture model,nonlinear dynamics,optical filters,tracking,mixture models,posterior distribution,graph theory,particle filters,predictive models,particle filter,belief propagation,gaussian distribution,bayesian methods | Computer vision,Nonlinear system,Pattern recognition,Computer science,Particle filter,Posterior probability,Approximate inference,Artificial intelligence,Motion estimation,Probabilistic logic,Mixture model,Computation | Conference |
ISSN | Citations | PageRank |
1063-6919 | 13 | 1.75 |
References | Authors | |
20 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Oscar Nestares | 1 | 134 | 12.37 |
David J. Fleet | 2 | 5236 | 550.74 |