Presentation 2003/1/28
Automatic Detection of Perceptually Important Regions in a Color Image
Takasi YAHARA, Junji MAEDA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Humans generally focus on a few areas in an image rather than scan all areas of a scene. Extraction of perceptully important regions is useful for a wide applications such as object recognition, image compression and machine vision. We present a method to automatically produce an importance map, that predicts the perceptual importance for each segmented region in the image, using various factors which influence human visual attention. We used region growing algorithm as image segmentation technique. By using fuzzzy inference to develop an importance map, we obtained the results that coincide with human sence.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Image Segmentation / Region Growing / Importance Map / Fuzzy Inference
Paper # ITS2002-94,IE2002-235
Date of Issue

Conference Information
Committee IE
Conference Date 2003/1/28(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 Image Engineering (IE)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Automatic Detection of Perceptually Important Regions in a Color Image
Sub Title (in English)
Keyword(1) Image Segmentation
Keyword(2) Region Growing
Keyword(3) Importance Map
Keyword(4) Fuzzy Inference
1st Author's Name Takasi YAHARA
1st Author's Affiliation Department of Computer Science and Systems Engineering, Faculty of Enginnering, Muroran Institute of technology()
2nd Author's Name Junji MAEDA
2nd Author's Affiliation Department of Computer Science and Systems Engineering, Faculty of Enginnering, Muroran Institute of technology
Date 2003/1/28
Paper # ITS2002-94,IE2002-235
Volume (vol) vol.102
Number (no) 632
Page pp.pp.-
#Pages 5
Date of Issue