A Novel Discriminant Non-Negative Matrix Factorization Algorithm With Applications to Facial Image Characterization Problems

The methods introduced so far regarding discriminant non-negative matrix factorization (DNMF) do not guarantee convergence to a stationary limit point. In order to remedy this limitation, a novel DNMF method is presented that uses projected gradients. The proposed algorithm employs some extra modifi...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on information forensics and security Vol. 2; no. 3; pp. 588 - 595
Main Authors: Kotsia, I., Zafeiriou, S., Pitas, I.
Format: Journal Article
Language:English
Published: New York IEEE 01.09.2007
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Algorithms
Classification
Convergence
Eyebrows
Eyes
Face recognition
Facial
Facial expression recognition
Facial features
Factorization
frontal face verification
Image databases
Image recognition
Informatics
Linear discriminant analysis
non-negative matrix factorization (NMF)
Principal component analysis
projected gradients
Tasks
ISSN:1556-6013, 1556-6021
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The methods introduced so far regarding discriminant non-negative matrix factorization (DNMF) do not guarantee convergence to a stationary limit point. In order to remedy this limitation, a novel DNMF method is presented that uses projected gradients. The proposed algorithm employs some extra modifications that make the method more suitable for classification tasks. The usefulness of the proposed technique to frontal face verification and facial expression recognition problems is demonstrated.
Bibliography:ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
content type line 23
ISSN:1556-6013
1556-6021
DOI:10.1109/TIFS.2007.902017