Presentation 2008-01-17
Attention Propagation for Object Detection
Seihiro Sakahira, Toshikazu Wada,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Visual object detection is one of the most basic technology for object tracking and recognition. Existing methods performes detection for independent image frames. In the case of human vision system, however, exploits the previous detection results, i.e., detection will not be performed within the regions where no object has been detected in the previous image frame, and will be performed within other regions. This report proposes a detection framework that exploits previous detection results by propagating the potential values representing how much shall the detection pay attention. The algorithm for maintaining this potential values and for detecting the objects by referring the potential values are proposed and examined through the face detection experiments.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) AdaBoost / Cascade classifiers / Haar features
Paper # PRMU2007-173
Date of Issue

Conference Information
Committee PRMU
Conference Date 2008/1/10(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 Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Attention Propagation for Object Detection
Sub Title (in English)
Keyword(1) AdaBoost
Keyword(2) Cascade classifiers
Keyword(3) Haar features
1st Author's Name Seihiro Sakahira
1st Author's Affiliation Graduate School of Systems Engineering, Wakayama University()
2nd Author's Name Toshikazu Wada
2nd Author's Affiliation Graduate School of Systems Engineering, Wakayama University
Date 2008-01-17
Paper # PRMU2007-173
Volume (vol) vol.107
Number (no) 427
Page pp.pp.-
#Pages 6
Date of Issue