An efficient classifier fusion for face recognition including varying illumination

Authors:
Mi Young Nam;Jo Hyung Yoo;Phill Kyu Rhee
Affiliations:
Dept. of Computer Science & Engineering, Inha University, Incheon, South Korea;Dept. of Computer Science & Engineering, Inha University, Incheon, South Korea;Dept. of Computer Science & Engineering, Inha University, Incheon, South Korea
Venue:
PCM'05 Proceedings of the 6th Pacific-Rim conference on Advances in Multimedia Information Processing - Volume Part II
Year:
2005

Citing 4
Cited 0

Optimal combinations of pattern classifiers

Pattern Recognition Letters
On Combining Classifiers

IEEE Transactions on Pattern Analysis and Machine Intelligence
Genetic Algorithms in Search, Optimization and Machine Learning

Genetic Algorithms in Search, Optimization and Machine Learning
Fuzzy ARTMAP network with evolutionary learning

ICASSP '00 Proceedings of the Acoustics, Speech, and Signal Processing, 2000. on IEEE International Conference - Volume 06

Quantified Score

Hi-index	0.02

Visualization

Abstract

In this paper, we propose an efficient classifier fusion for face recognition under varying illumination environments by taking classifier fusion's advantage of environment context identification. Adaptation to dynamically changing environments is very important since advanced applications become pervasive and ubiquitous. The proposed classifier fusion system, called BCF (Bayesian based classifier fusion), adopts the concept of face context awareness and evolutionary computing. But aside the difference of classifiers the training data performs main role in them consequently the results from the classifiers couldn't be so individual from each other to make decision by considering them. The system working environments are clustered and identified as face environmental context. Majority voting (MV), Maximum based fusion (MX) and Minimum based fusion (MN) are used to explore the most effective action configuration for each identified context. Once the context knowledge is constructed, the system can adapt to varying environment in real-time. The superiority of the proposed scheme is shown using three face image data sets: Inha, FERET, and Yale.