儘管支持向量機（Support Vector Machine, SVM）是屬於強分類器，不容易與調昇式學習法（Adaptive Boosting, AdaBoost）作結合，但其已被證明為調昇式學習法中有效的組合分類器（Component Classifier）。本論文即根據此理論提出以特徵篩選的調昇式學習法結合支持向量機來實作性別和年齡辨識。所選用的特徵皆被證實能有效描述人臉，包含局部二元化模版（Local Binary Pattern, LBP）、局部方向性模版（Local Directional Pattern, LDP）、局部三元化模版（Local Ternary Pattern, LTP）、蓋伯濾波器（Gabor Filter）以及局部蓋伯二元化模版（Local Gabor Binary, LGBP）。上述各種方式擷取的特徵，都具有很高的維度，為了加速模型的訓練，以及效能評估的檢測，我們另外使用兩種方式進行降維：主成分分析（Principal Component Analysis, PCA）以及線性識別分析（Linear Discriminant Analysis, LDA），接著利用支持向量機作為AdaBoost的組合分類器，即可在AdaBoost每次迭代的過程篩選出最適合的特徵，進而學習並組合出最終的辨識模型。
由辨識結果看來，未進行特徵篩選的AdaBoostSVM與SVM這兩個分類器有近似的辨識能力，而進行特徵篩選後的AdaBoostSVM則有機會得到較上述兩分類器好的性別和年齡辨識結果。

Although SVM (Support Vector Machine) is a strong classifier that cannot be combined with AdaBoost (Adaptive Boosting) easily, it has been proved to be effective component classifier in AdaBoost. In this paper, we propose AdaBoost multiple feature selection with SVM for gender and age classification based on the above-mentioned theory. The chosen features have been proved to be successful in describing facial image. They include LBP (Local Binary Pattern), LDP (Local Directional Pattern), LTP (Local Ternary Pattern), Gabor filter, and LGBP (Local Gabor Binary Pattern). To speed up the training process and get better performance, we reduce dimensionalities by two methods: PCA (Principal Component Analysis) and LDA (Linear Discriminant Analysis). Then, we use SVM as the component classifier in AdaBoost. At each iteration in AdaBoost, we can choose the most suitable feature by evaluating the training error rate. Finally, we can combine these component classifiers into the ultimate recognition model.
From the experimental results, AdaBoostSVM with no feature selection can perform as well as SVM. Moreover, AdaBoostSVM with multiple feature selection has the chance to outperform the other classifiers in both gender and age classification.

摘要 I
Abstract II
謝誌 III
目錄 IV
圖目次 VII
表目次 X
第一章 緒論 1
1.1 研究動機 1
1.2 性別與年齡辨識系統 1
1.3 研究方向 2
1.4 章節概要 3
第二章 相關研究簡介 4
第三章 研究方法 6
3.1 影像前處理 6
3.1.1 人臉偵測 6
3.1.2 人臉校準 7
3.1.3 大小調整、灰階化、直方圖均化 8
3.2 特徵擷取 8
3.2.1 灰階強度特徵 8
3.2.2 局部二元化模版 9
3.2.3 局部方向性模版 11
3.2.4 局部三元化模版 13
3.2.5 蓋伯濾波器 14
3.2.6 局部蓋伯二元化模版 15
3.3 降低特徵維度 16
3.3.1 減少取樣點 16
3.3.2 主成分分析 17
3.3.3 線性識別分析 17
3.4 辨識模型訓練 18
3.4.1 支持向量機 18
3.4.2 調昇式學習法 19
3.4.3 調昇式學習法結合支持向量機 21
第四章 研究結果與分析 24
4.1 資料庫簡介 24
4.1.1 PAL 人臉資料庫 24
4.1.2 FERET 人臉資料庫 26
4.1.3 MIR-Google 人臉資料庫 27
4.2 實驗參數設定 29
4.3 效能評估方式 30
4.4 性別辨識實驗 31
4.4.1 PAL 人臉資料庫實驗 31
4.4.2 FERET 人臉資料庫實驗 33
4.4.3 MIR-Google 人臉資料庫實驗 34
4.5 年齡辨識實驗 36
4.5.1 PAL 人臉資料庫實驗 37
4.5.2 FERET 人臉資料庫實驗 39
4.5.3 MIR-Google 人臉資料庫實驗 40
4.6 實驗錯誤分析 43
4.6.1 性別辨識錯誤分析 43
4.6.2 年齡辨識錯誤分析 44
第五章 結論與未來研究方向 46
參考文獻 47

[1] 第八屆由田機器視覺獎（Utechzone Machine Vision Prize, UTMVP）http://www.utmvp.com/
[2] P. Jonathon Phillips, Harry Wechsler, Jeffrey Huang, Patrick J. Rauss: The FERET database and evaluation procedure for face-recognition algorithms. Image Vision Comput. 16(5): 295-306 (1998).
[3] Minear, Meredith, and Denise C. Park. "A lifespan database of adult facial stimuli." Behavior Research Methods, Instruments, & Computers 36.4 (2004): 630-633.
[4] Lian, Hui-Cheng, and Bao-Liang Lu. "Multi-view gender classification using local binary patterns and support vector machines." Advances in Neural Networks-ISNN 2006. Springer Berlin Heidelberg, 2006. 202-209
[5] Xia, Bin, He Sun, and Bao-Liang Lu. "Multi-view gender classification based on local Gabor binary mapping pattern and support vector machines." Neural Networks, 2008. IJCNN 2008.(IEEE World Congress on Computational Intelligence). IEEE International Joint Conference on. IEEE, 2008
[6] Jabid, Taskeed, M. Hasanul Kabir, and Oksam Chae. "Gender classification using local directional pattern (LDP)." Pattern Recognition (ICPR), 2010 20th International Conference on. IEEE, 2010.
[7] Gunay, Asuman, and Vasif V. Nabiyev. "Automatic age classification with LBP." Computer and Information Sciences, 2008. ISCIS'08. 23rd International Symposium on. IEEE, 2008.
[8] Gao, Feng, and Haizhou Ai. "Face age classification on consumer images with gabor feature and fuzzy lda method." Advances in biometrics. Springer Berlin Heidelberg, 2009. 132-141.
[9] Shan, Caifeng. "Learning local features for age estimation on real-life faces." Proceedings of the 1st ACM international workshop on Multimodal pervasive video analysis. ACM, 2010.
[10] Li, Xuchun, Lei Wang, and Eric Sung. "AdaBoost with SVM-based component classifiers." Engineering Applications of Artificial Intelligence 21.5 (2008): 785-795.
[11] Viola, Paul, and Michael Jones. "Rapid object detection using a boosted cascade of simple features." Computer Vision and Pattern Recognition, 2001. CVPR 2001. Proceedings of the 2001 IEEE Computer Society Conference on. Vol. 1. IEEE, 2001.
[12] Shih, Frank Y., and Chao-Fa Chuang. "Automatic extraction of head and face boundaries and facial features." Information Sciences 158 (2004): 117-130.
[13] Ojala, Timo, Matti Pietikäinen, and David Harwood. "A comparative study of texture measures with classification based on featured distributions." Pattern recognition 29.1 (1996): 51-59.
[14] T. Ahonen, A. Hadid, and M. Pietikainen, “Face Description with Local Binary Patterns: Application to Face Recognition,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 12, pp. 2037-2041, Dec. 2006.
[15] Shan, Caifeng. "Learning local features for age estimation on real-life faces." Proceedings of the 1st ACM international workshop on Multimodal pervasive video analysis. ACM, 2010.
[16] Tan, Xiaoyang, and Bill Triggs. "Enhanced local texture feature sets for face recognition under difficult lighting conditions." Image Processing, IEEE Transactions on 19.6 (2010): 1635-1650.
[17] Luu, Khoa, et al. "Combined local and holistic facial features for age-determination." Control Automation Robotics & Vision (ICARCV), 2010 11th International Conference on. IEEE, 2010.
[18] Lee, Tai Sing. "Image representation using 2D Gabor wavelets." Pattern Analysis and Machine Intelligence, IEEE Transactions on 18.10 (1996): 959-971.
[19] Zhang, Wenchao, et al. "Local Gabor binary pattern histogram sequence (LGBPHS): A novel non-statistical model for face representation and recognition." Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on. Vol. 1. IEEE, 2005.
[20] Jolliffe, Ian T. Principal component analysis. Vol. 487. New York: Springer-Verlag, 1986.
[21] Deng, Hong-Bo, et al. "A new facial expression recognition method based on local gabor filter bank and pca plus lda." International Journal of Information Technology 11.11 (2005): 86-96.
[22] Liu, Chengjun. "Gabor-based kernel PCA with fractional power polynomial models for face recognition." Pattern Analysis and Machine Intelligence, IEEE Transactions on 26.5 (2004): 572-581.
[23] Mika, Sebastian, et al. "Fisher discriminant analysis with kernels." Neural Networks for Signal Processing IX, 1999. Proceedings of the 1999 IEEE Signal Processing Society Workshop.. IEEE, 1999.
[24] Cortes, Corinna, and Vladimir Vapnik. "Support-vector networks." Machine learning 20.3 (1995): 273-297
[25] Freund, Yoav, and Robert E. Schapire. "Experiments with a new boosting algorithm." ICML. Vol. 96. 1996.
[26] Chang, Chih-Chung, and Chih-Jen Lin. "LIBSVM: a library for support vector machines." ACM Transactions on Intelligent Systems and Technology (TIST)2.3 (2011): 27.
[27] Ojala, Timo, Matti Pietikainen, and Topi Maenpaa. "Multiresolution gray-scale and rotation invariant texture classification with local binary patterns." Pattern Analysis and Machine Intelligence, IEEE Transactions on 24.7 (2002): 971-987.
[28] Liao, Wen-Hung. "Region description using extended local ternary patterns."Pattern Recognition (ICPR), 2010 20th International Conference on. IEEE, 2010.