Performance Evaluation of Face Recognitions
In this paper, we make comparisons on ORL and FERET face database for several face recognition algorithms, which are based on PCA, LI-PCA, LDA, RLDA, NMF and LNMF. The performance evaluations of accuracy and stability of the algorithms are given. The subset bootstrap algorithm is suitable for face samples which dont meet the condition of independent identically distributed. For the stability evaluation of algorithms, with the confidence interval of equal error rate being the stability index, the subset bootstrap technique is used to evaluate the stability performance of the algorithms. Experiment results show that the subset bootstrap method gives accurate evaluation of stability performance. The face recognition based on RLDA shows the best robustness with the smallest EER, but the best stable one is traditional LDA.
face recognition PCA LDA NMF performance evaluation subset bootstrap
Heng Zhao Jia Cui Ding Linghu
Schoool of Life Sciences and Technology, Xidian University Xian, Shaanxi, China School of Electronic Engineering, Xidian University , Xian, Shaanxi, China
国际会议
哈尔滨
英文
317-321
2012-05-19（万方平台首次上网日期，不代表论文的发表时间）