由於傳統區別分析中常用 PCA 進行前處理，但 PCA 前處理可
能破壞原始資料的結構，使區別能力下降，為了降低 PCA 的負面影
響，本篇論文提出將原空間的資料透過不同的 kernel Methods 映射至
kernel 空間，目的是使資料在特徵空間中更具區別性，再於特徵空間
使用 PCA 將主成分提出；並配合最近特徵線轉換法 (NFLE) 進行區
別分析，以提昇辨識率。
在實驗部分，我們使用 CMU 資料庫做人臉辨識，針對不同的
樣本數與不同的維度下進行效能驗證 NFLE , Linear Kernel+NFLE ,
Guassian Kernel+NFLE , Polynomial Kernel +NFLE 四種不同的演算
法；
由實驗數據可得，雖然 Kernel Methods 在低樣本數的辨識率低
於 NFLE，但隨著樣本數增多，Kernel Methods 的辨識率則會高於
NFLE。

In traditional discriminant analysis, PCA is usually applied for data
preprocessing. However, PCA may bring damage to the topology of
original data and hence decrease the discriminability. To remedy this
problem, the Kernel methods are adopted to transform the data set
from original space to feature space for enhancing the discriminability
in this study. In the kernel space, the PCA is then applied to extract
the principal component data and remove the noises. After the PCA
process, the NFLE algorithm is applied for discriminant analysis.
In the experiments, Linear Kernel+NFLE , Guassian Kernel+NFLE,
Polynomial Kernel+NFLE algorithms are implemented for face recognition. In our work,the CMU face database is used for evaluating the
performance of the proposed methods in different training samples and
dimensions.
Experimental results reveal that the recognition rate of the proposed kernel based method is lower than NFLE under few training samples. When the number of training samples increases, the proposed
kernel based method outperforms NFLE.

[1] 陳順東. 「虹膜辨識系統之研究與實作」. Master’s thesis, 國立中
央大學, 2004.
[2] 鄭仁和. 「基於色彩資訊特徵之虹膜辨識」. Master’s thesis, 國立
暨南國際大學, 2009.
[3] 黃玟瑜. 「使用影像處理技術於糖尿病視網膜病變之自動化辨識」
. Master’s thesis, 國立嘉義大學, 2008.
[4] 薛傑仁. 「生物辨識之人臉辨識的方法」. Master’s thesis, 亞洲大
學, 2010.
[5] 王俐芃. 「3d 影像應用於人臉辨識之實現」. Master’s thesis, 國立
臺北教育大學, 2011.
[6] L. Sirovich and M. Kirby. “low-dimensional procedure for the
characterization for the characterization of human faces”. J. Opt.
Soc. Am. A 4, 519-524, 1987.
[7] M. A. Turk and A. P. Pentland. “face recognition using eigenfaces” .
Computer Vision and Pattern Recognition, 1991. Proceedings CVPR
’91., pages 586 – 591, 1991.
[8] 张晓红、汤晓华、沈晓红. 「基于 kernel pca 的人脸识别算法的探
讨」. 北京工商大学学报 (自然科学版), 2008.
[9] 徐茂翔. 「人臉偵測與基於鑑別性特徵之超暗人臉辨識」 . Master’s
thesis, 大葉大學, 2011.
[10] Bernhard Schölkopf, Alexander Smola, and Klaus-Robert Müller.
“nonlinear component analysis as a kernel eigenvalue problem”.
Neural Computation, 10:1299–1319, 1998.
[11] B. Schölkopf, A. Smola, and K.-R. Müller. “kernel principal component analysis”. Advances in Kernel Methods—Support Vector
Learning, pages 327–352, 1999.
[12] Quan Wang. “kernel principal component analysis and its applications in face recognition and active shape models”. CoRR, abs/
1207.3538, 2012.
[13] Kilian Q. Weinberger, Fei Sha, and Lawrence K. Saul. “learning a
kernel matrix for nonlinear dimensionality reduction”. In In Proceedings of the Twenty First International Conference on Machine
Learning (ICML-04), pages 839–846. ACM Press, 2004.
[14] Roni Khardon. “more kernels and their properties”.
取自 http://www.cs.tufts.edu/~roni/Teaching/CLT2008S/LN/
lecture18.pdf, 2008.
[15] Wen Ge, Xu Hongzhe, Zheng Weibin, Zhong Weilu, and Fu Baiyang.
“multi-kernel pca based high-dimensional images feature reduction”
. Electric Information and Control Engineering (ICEICE), pages
5966 – 5969, 2011.
[16] 韓欽銓、陳映濃. 「最近特徵空間轉換法之人臉辨識」. 取
自 http://vbie.eic.nctu.edu.tw/vol_37/index_firefox_
frontpage_epaper.htm#%B3%CC%AA%F1%AFS%BCx%AA%C5%B6%A1%
C2%E0%B4%AB%AAk%A4%A7%A4H%C1y%BF%EB%C3%D1, 7 月 2010.
[17] 陳映濃. 「應用最近特徵空間轉換法於人臉辨識」. PhD thesis, 中
央大學, 2011.
[18] 丁維雷、付永慶. 「kpca 特徵提取算法的自動目標識別」. 應用科
技, 2011/9.
[19] 陳映濃、何岡峰、韓欽銓、范國清. 「多階段人臉偵測與辨識」.
前瞻科技與管理, pages 47–61, 2013/11.
[20] Terence Sim, Simon Baker, and Maan Bsat. “the cmu pose, illumination, and expression (pie) database”, 2002.