- Abstract:
-
We investigate a solution to the problem of multi-sensor perception by formulating it in the framework of Bayesian model selection. Humans robustly integrate and segregate multi-sensory data as appropriate, but previous theoretical work has focused largely on purely integrative cases, leaving segregation unaccounted for and unexploited by machine perception systems. We illustrate a unifying, principled Bayesian solution which accounts for both integration and segregation by reasoning explicitly about data association in a probabilistic framework. Unsupervised learning of such a model with EM is illustrated for a real world audio-visual application.
- Links To Paper
- 1st Link
- Bibtex format
- @Misc{EDI-INF-RR-0829,
- author = {
Timothy Hospedales
and Sethu Vijayakumar
},
- title = {Bayesian multisensory perception},
- year = 2006,
- month = {Jun},
- url = {http://www.ipab.inf.ed.ac.uk/slmc/SLMCpeople/Hospedales_T_files/tr-bms.pdf},
- }
|