Illumination invariant face image representation using quaternions

Authors:
Dayron Rizo-Rodríguez;Heydi Méndez-Vázquez;Edel García-Reyes
Affiliations:
Advanced Technologies Application Center, Havana, Cuba;Advanced Technologies Application Center, Havana, Cuba;Advanced Technologies Application Center, Havana, Cuba
Venue:
CIARP'10 Proceedings of the 15th Iberoamerican congress conference on Progress in pattern recognition, image analysis, computer vision, and applications
Year:
2010

Citing 6
Cited 2

Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection

IEEE Transactions on Pattern Analysis and Machine Intelligence
Face Recognition Using the Discrete Cosine Transform

International Journal of Computer Vision - Special issue: Research at McGill University
Acquiring Linear Subspaces for Face Recognition under Variable Lighting

IEEE Transactions on Pattern Analysis and Machine Intelligence
An Illumination Insensitive Representation for Face Verification in the Frequency Domain

ICPR '06 Proceedings of the 18th International Conference on Pattern Recognition - Volume 01
"Corefaces"- robust shift invariant PCA based correlation filter for illumination tolerant face recognition

CVPR'04 Proceedings of the 2004 IEEE computer society conference on Computer vision and pattern recognition
Hypercomplex correlation techniques for vector images

IEEE Transactions on Signal Processing

Quaternion correlation filters for illumination invariant face recognition

CIARP'11 Proceedings of the 16th Iberoamerican Congress conference on Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications
Illumination Invariant Face Recognition Using Quaternion-Based Correlation Filters

Journal of Mathematical Imaging and Vision

Quantified Score

Hi-index	0.00

Visualization

Abstract

Variations in illumination is a well-known affecting factor of face recognition system performance. Feature extraction is one of the principal steps on a face recognition framework, where it is possible to alleviate the illumination effects on face images. The aim of this work is to study the illumination invariant properties of a hypercomplex image representation. A quaternion description from the image is built using second order derivatives decomposition. This representation is transformed to quaternion frequency domain in order to analyze its illumination invariant and discriminative properties, which are compared against the ones of the complex frequency domain representation obtained by using first order derivative decomposition. The hypercomplex quaternion representation was found to be more discriminative than the complex one, when comparing on face recognition with images under varying lighting conditions.