- Abstract:
-
The research proposed in this document will seek to investigate the problem in Computer Vision of tracking objects during encounters and interaction in a scene monitored by a single static camera. Humans do this exceptionally well because we have very high resolution stereo vision and a well calibrated low-level vision system to identify and separate objects, even when some are partially or fully occluded. We also rely heavily on strong priming from our higher-level reasoning system to explain what is going on in the scene, and we base this on a continuous analysis of current and recent events and on expectations about the scene estimated from context as well as on perception of the 3D world. We have a unique ability to re-evaluate a situation based on new evidence and re-estimate previous assumptions based on current probability calculations.
Computers have in comparison very limited sensors and an even more limited ability to reason about their environment. Using Computer Vision they can track separate individuals in a scene quite well (Crowley and Reignier, 2003) and some research has shown that by keeping track of the characteristics of individuals it is possible to reacquire a tracked target after an encounter or occlusion with quite good results (Jorge, Abrantes and Marques, 2004), (Chomat, Verdire, Hall and Crowley, 2000), (Jorge, Abrantes and Marques, 2004).
Solving the problem of tracking individuals or members of identified groups during encounters has been very sparsely addressed and is usually avoided completely by merely creating a new multi-object individual (Mabey and Gunther, 2003) or by disregarding most of the collected information during the encounter as random noise.
We hope to mimic the human approach by collecting detailed information from each moment and keeping this in hierarchical spatio-temporal structures, which can then be traversed, analysed and modified by a higher-level reasoning system as new information becomes availa
- Links To Paper
- 1st link
- Bibtex format
- @Misc{EDI-INF-RR-0266,
- author = {
Thor List
},
- title = {Thesis Proposal: Tracking Close Proximity Encounters},
- year = 2005,
- month = {Mar},
- url = {http://www.ipab.inf.ed.ac.uk/PROPOSALS/list_thesis_proposal.pdf},
- }
|