Title :
Dynamically optimizing face recognition using PCA
Author :
Zhihong Zhao; Beibei Li
Author_Institution :
Electronic and automation department, Nanjing Institute of Industry Technology, China
Abstract :
A dynamically optimizing PCA method for face recognition is presented. Firstly, DCT is adopted, and their mean gray values are obtained. The face images are grouped into three sorted groups: Large, Medium and Small, and then calculate their means. Then, the Eigenvalue number is chosen as 9 and 10. Finally, recognition and reconstruction may be accomplished. The results show that this method achieved better results compared to a number of state of the art PCA variants.
Keywords :
"Principal component analysis","Face","Databases","Discrete cosine transforms","Face recognition","Frequency-domain analysis","Covariance matrices"
Conference_Titel :
Control, Automation and Information Sciences (ICCAIS), 2015 International Conference on
DOI :
10.1109/ICCAIS.2015.7338646
