Presentation 2005/1/17
Saliency-based visual search : how to code and utlize visual sailency
Takahiko KOIKE, Jun SAIKI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In order to investigate the mechanism of visual attetnion, visual search tasks are used in experimental psychology. A saliency-based visual attention model can simulate human behavior, i.e. efficient search and inefficient search in visual search tasks. A saliency-based model assumes that visual saliency is statically coded on a saliency map, and that the rank order of directing attention is identical to the rank order of visual saliency. However, it is not clear that the saliency-based mehcnaism is best suited to account for human behavior in visual search tasks. In this study, we proposed a computational model for a saliency-based visual attention model. Using the model, we investigated how information of visual sailency on a saliency map is coded and used by a visual attention mechanism. Our results suggest that a coding of visual saliency may fluctuate substantially, thus the rank order of shifting attentional focus may be not same to the rank ordr of saliency. The simulation also showed the possibility that an inhibitory tagging mechanism does not make visual search more efficient by prohibiting an attention from re-directing to previously attended location.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) visual attention / visual search task / visual saliency / inhibitory tagging
Paper # NC2004-116
Date of Issue

Conference Information
Committee NC
Conference Date 2005/1/17(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Saliency-based visual search : how to code and utlize visual sailency
Sub Title (in English)
Keyword(1) visual attention
Keyword(2) visual search task
Keyword(3) visual saliency
Keyword(4) inhibitory tagging
1st Author's Name Takahiko KOIKE
1st Author's Affiliation Graduate School of Informatics, Kyoto University()
2nd Author's Name Jun SAIKI
2nd Author's Affiliation Japanese Science and Technology Agency:Graduate School of Informatics, Kyoto University
Date 2005/1/17
Paper # NC2004-116
Volume (vol) vol.104
Number (no) 585
Page pp.pp.-
#Pages 6
Date of Issue