Presentation 1996/2/3
Reduction of Computational Cost in Object Recognition using GRBF Network
Mitsuhiro SEKITO, Kunikazu KOBAYASHl, Toyoshi TORlOKA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) It is reported that GRBF networks can recognize 3D flexible objects from their 2D view. The network recognizes 3D flexible wire-frame objects by approximating the mapping from an arbitrary view to the standard view by learning. The former recognition method is using the feature point matching based on characteristic function and geometric condition. Therefore, it needs lots of computational cost. In this paper, we propose a new recognition method which uses a sorting without the matching. Our method can remarkably reduce computational cost and apply to 3D objects with occluded parts in their 2D view.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) GRBF / 3D object recognition / sorting / backpropergation
Paper # NC95-101
Date of Issue

Conference Information
Committee NC
Conference Date 1996/2/3(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) Reduction of Computational Cost in Object Recognition using GRBF Network
Sub Title (in English)
Keyword(1) GRBF
Keyword(2) 3D object recognition
Keyword(3) sorting
Keyword(4) backpropergation
1st Author's Name Mitsuhiro SEKITO
1st Author's Affiliation Fac. of Engineering, Yamaguchi University()
2nd Author's Name Kunikazu KOBAYASHl
2nd Author's Affiliation Fac. of Engineering, Yamaguchi University
3rd Author's Name Toyoshi TORlOKA
3rd Author's Affiliation Fac. of Engineering, Yamaguchi University
Date 1996/2/3
Paper # NC95-101
Volume (vol) vol.95
Number (no) 506
Page pp.pp.-
#Pages 6
Date of Issue