Paper Info
Title
The Face Recognition Algorithm Based on Offset Difference of Double Subspace
Abstract
In subspace approaches for the pattern recognition, the transform fashion is paid more attentions but the correlation between subspaces is given little concern in previous research. By mapping the space of all training samples to the corresponding subspace of individual training sample using PCA, we discover that there is a very strong relationship between two subspaces. Specially, a higher mutual compensability and consistency appears in both of these two subspaces. Therefore, a new recognition algorithm based on the difference of double subspaces is presented in this paper. The new algorithm sufficiently utilizes the relativity of PCA eigen-subspaces of the total sample and individual sample spaces of the sample to be recognized, so that it improve efficiently the recognition rate. We prove the validity of the proposed algorithm under some mild divisible condition, and give some the experiments to demonstrate that the new algorithm has higher recognition rate than some similar algorithms.
Year
DOI
Venue
2009
10.1109/FSKD.2009.233
FSKD (1)
Keywords
DocType
Citations 
similar algorithm,pattern recognition,higher recognition rate,individual sample space,new algorithm,proposed algorithm,face recognition algorithm,new recognition algorithm,double subspaces,offset difference,recognition rate,individual training sample,algorithm design and analysis,prototypes,face,face recognition,principal component analysis,pca,databases
Conference
0
PageRank 
References 
Authors
0.34
12
2
Name
Order
Citations
PageRank
Shibin Xuan100.68
LeJun Shen2151.15