Presentation 2000/9/15
Refinements of Virtual Subspace Method for Robust Face Recognition in Natural Lighting Conditions
Kazuma Shigenari, Takeshi Shakunaga,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The virtual subspace method was proposed in [7] for the robust face recognition independent of lighting conditions. This paper refines the original method in order to reduce computational costs and to achieve a high recognition rate. Some refinements are made with the original method by (1) introduction of the normalized correlation as the similarity measure, (2) construction of the virtual subspace in the canonical space, and (3) decomposition of the similarity measure into two orthogonal factors. Final results show that 89.7% correct discrimination is made over 50 persons when only one image is registered for each person.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Virtual subspace method / face recognition / canonical eigenspace / lighting compensation / normalized correlation / projection residual
Paper # PRMU2000-79
Date of Issue

Conference Information
Committee PRMU
Conference Date 2000/9/15(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) Refinements of Virtual Subspace Method for Robust Face Recognition in Natural Lighting Conditions
Sub Title (in English)
Keyword(1) Virtual subspace method
Keyword(2) face recognition
Keyword(3) canonical eigenspace
Keyword(4) lighting compensation
Keyword(5) normalized correlation
Keyword(6) projection residual
1st Author's Name Kazuma Shigenari
1st Author's Affiliation Department of Information Technology, Okayama University()
2nd Author's Name Takeshi Shakunaga
2nd Author's Affiliation Department of Information Technology, Okayama University
Date 2000/9/15
Paper # PRMU2000-79
Volume (vol) vol.100
Number (no) 312
Page pp.pp.-
#Pages 8
Date of Issue