Using Quotient Image and Linear Subspace for Face Recognition under Arbitrary Illumination Conditions

Authors:
Lin He;Jiwei Liu;Bo Zhang;Zhiliang Wang
Affiliations:
-;-;-;-
Venue:
CAR '09 Proceedings of the 2009 International Asia Conference on Informatics in Control, Automation and Robotics
Year:
2009

Citing 7
Cited 0

Face Recognition: The Problem of Compensating for Changes in Illumination Direction

IEEE Transactions on Pattern Analysis and Machine Intelligence
The Quotient Image: Class-Based Re-Rendering and Recognition with Varying Illuminations

IEEE Transactions on Pattern Analysis and Machine Intelligence
From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose

IEEE Transactions on Pattern Analysis and Machine Intelligence
Analytic PCA Construction for Theoretical Analysis of Lighting Variability in Images of a Lambertian Object

IEEE Transactions on Pattern Analysis and Machine Intelligence
Lambertian Reflectance and Linear Subspaces

IEEE Transactions on Pattern Analysis and Machine Intelligence
Face Recognition from a Single Training Image under Arbitrary Unknown Lighting Using Spherical Harmonics

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Theory Of Frequency Domain Invariants: Spherical Harmonic Identities for BRDF/Lighting Transfer and Image Consistency

IEEE Transactions on Pattern Analysis and Machine Intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we theoretically analyze the limitation of the Quotient Image (QI) method proposed by Shashua and present a new way to compute the quotient image. Based on the observation that nine basis point light sources can represent almost arbitrary lighting conditions for face recognition application, we use the corresponding nine basis images which are all real images to span the image space. And similar to the QI, our method can not only get an illumination invariant image but also synthesize a new image under arbitrary illumination conditions. We provide an experimental result to show the effectiveness of our algorithm as well.