| ECVision indexed and annotated bibliography of cognitive computer vision publications This bibliography was created by Hilary Buxton and Benoit Gaillard, University of Sussex, as part of ECVision Specific Action 8-1 The complete text version of this BibTeX file is available here: ECVision_bibliography.bib
H. D. Tagare and K. Toyama and J. G. Wang
ABSTRACT
A precise analysis of an entire image is computationally wasteful if one is interested in finding a target object that is located in a sub-region of the image. A useful attention strategy can reduce the overall computation by carrying out fast but approximate image measurements and using their results to suggest promising regions of the image. This paper proposes a maximum-likelihood attention strategy which uses the results of a fast pre-attentive module to direct a slower but more precise post-attentive module to those regions of the image that are likely to contain the target. The attention mechanism recognizes that objects are made of parts and that parts have di erent features. It works by proposing object part-image feature pairings which have the highest likelihood of coming from the target. The exact calculation of the likelihood as well as approximations are provided. The attention mechanism is adaptive, that its behavior depends on the statistics of the image features. It also exhibits \pop-out{"} and \camou which are behaviors exhibited by human visual attention. Experimental results suggest that, on average, the algorithm evaluates less than 2 of all part-feature pairs before selecting the actual object, showing a significant reduction in the complexity of visual search. |
|
| Site generated on Friday, 06 January 2006
|