Basic Search / Detailed Display

Author: 許傑貫
Chieh-Kuan Hsu
Thesis Title: 階層式年齡估測方法利用人臉特徵以及皺紋偵測之研究
A Study of Hierarchical Age Estimation Using Facial Features and Wrinkle Detection
Advisor: 吳怡樂
Yi-Leh Wu
Committee: 何瑁鎧
Maw-Kae Hor
唐政元
Cheng-Yuan Tang
鄧惟中
Wei-Chung Teng
Degree: 碩士
Master
Department: 電資學院 - 資訊工程系
Department of Computer Science and Information Engineering
Thesis Publication Year: 2011
Graduation Academic Year: 99
Language: 英文
Pages: 52
Keywords (in Chinese): 皺紋偵測主動外觀模型臉部特徵擷取支援向量機支援向量迴歸
Keywords (in other languages): Facial features extraction, Active Appearance Model, Wrinkle Detection, Support Vector Machine, Support Vector Regression
Reference times: Clicks: 281Downloads: 2
Share:
School Collection Retrieve National Library Collection Retrieve Error Report

人臉年齡的估測一直是一個不容易從外觀上判斷的問題。因為人的外觀會因為很多因素而有不同的變化,例如年齡的成長、不同人種之間的差異、性別的差異、生活環境、天氣等等因素都會有所不同。在年齡預測的研究中有兩個廣為人知的資料集分別是Face and Gesture Recognition Research Network (FG-NET) [18] 和Morphology (MORPH) [17],我們從中擷取部份的資料來進行實驗,實驗的對象在零到六十九歲之間。
基於Luu et al. [10] 提出的架構,以六種人臉特徵集合逐一實驗並與原始主動外觀模型 [20] 提供的68個點的結果做比較最後歸納出最佳的特徵集合。二十一歲到六十九歲年齡區間的實驗結果並不令人滿意,因此我們將年齡區間再分為二十一歲到四十歲以及四十一歲到六十九歲,並提出我們在年齡預估的方法。
基於我們的架構之下針對資料集1以及資料集2以六種人臉特徵點實驗,此架構首先由支援向量機 [4] 將影像資料分為三個年齡集合,分別是零到二十歲、二十一到四十歲、四十一到六十九歲,再根據支援向量迴歸 [8] 對上述三個年齡集合的每張測試影像預測迴歸值,最後根據此迴歸值去推估年齡。
我們也和其他優秀的年齡預估方法做比較,由實驗結果顯示我們的方法確實有助於改善人臉年齡推估。


Age estimation is not an easy problem from facial appearance. Because the facial appearance may be changed by many factors such as age, race, gender, living environment, weather, etc. There are two well established aging data sets, the Face and Gesture Recognition Research Network (FG-NET) [18] and the Morphology (MORPH) [17], for age estimation research. We utilize these two data sets and extract a part of them for our experiments with the subjects of age from 0 to 69 years old.
Based on Luu et al. [10] we implement six sets of facial features. We then compare each set of facial features with the Active Appearance Model (AAM) [20] using the original 68 features. From the result we conclude the sets of facial features that we will use. Furthermore, we find that the Mean Absolute Error (MAE) in the 21 to 69 years old age group is unsatisfactory. Hence we propose to divide this age group into 21 to 40 years old and 41 to 69 years old age groups and our age estimation architecture is also presented.
After that, we implement our architecture in six sets of facial features with dataset 1 and dataset 2. The image datasets are divided into three age groups which are 0 to 20 years old, 21 to 40 years old, and 41 to 69 years old by Support Vector Machine (SVM) [4] and get the regression value of each test image by Support Vector Regression (SVR) [8]. We then predict age by the prediction value of each test image. We also compare our result with other state-of-art age estimation methods. The experiment result shows that our propose method certainly improve the MAE in age estimation.

論文摘要.............................................I Abstract...........................................II List of Figures.....................................V List of Tables......................................VI 1. Introduction.....................................1 1.1 Motivation......................................1 1.2 Aging database..................................1 2. Related Work.....................................4 3. Age Estimation Methods...........................7 3.1 Active Appearance Model.........................7 3.1.1 AAM Training Process..........................7 3.1.2 AAM Fitting Process...........................8 3.2 Support Vector Machine..........................9 3.3 Support Vector Regression.......................11 3.4 Experiment with Different Facial Feature........12 3.4.1 The original 68 features in AAM...............12 3.4.2 Modified 49 features excluding lips...........14 3.4.3 Modified 32 features excluding eyebrows.......15 3.4.4 6 features of facial ratios...................17 3.4.5 Hybrid 21 features with facial ratios.........19 3.4.6 Hybrid 33 features with facial ratios.........21 3.4.7 Hybrid 36 features with wrinkle detection.....23 4. Experiments......................................28 4.1 Framework.......................................28 4.2 Experiment Result...............................29 5. Conclusion and Future Work.......................39 References .........................................40

[1]A. Lanitis, C. Draganova, and C. Christodoulou, “Comparing Different Classifiers for Automatic Age Estimation,” IEEE Trans. Systems, Man, and Cybernetics B, vol. 34, no. 1, pp. 621-628, Feb. 2004.
[2]A. Lanitis, C.J. Taylor, and T. Cootes, “Toward Automatic Simulation of Aging Effects on Face Images,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 4, pp. 442-455, Apr. 2002.
[3]B. Xiao, X. Yang, Y. Xu, and H. Zha, “Learning distance metric for regression by semidefinite programming with application to human age estimation,” Proceedings of the 17th ACM international conference on Multimedia, pp 451-460, Oct. 2009.
[4]C. Cortes and V. Vapnik, “Support-Vector Networks,” Machine Learning, vol. 20, no. 3, pp. 273-297, Sep. 1995.
[5]C.-C. Chang and C.-J. Lin, “LIBSVM: a library for support vector machines,”
Software available at http: //www.csie.ntu.edu.tw/˜cjlin/libsvm, 2001.
[6]C. Chen, Y. Chang, K. Ricanek, and Y. Wang, “Face Age Estimation Using Model Selection,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, pp. 93-99, Jun. 2010.
[7]G. Guo, Y. Fu, C. R. Dyer, and T. S. Huang, “Image-based human age estimation by manifold learning and locally adjusted robust regression,” IEEE Transactions on Image Processing, vol. 17, no. 7, pp. 1178-1188, Jul. 2008.
[8]H. Drucker, C. J.C. Burges, L. Kaufman, A. Smola and V. Vapnik. “Support Vector Regression Machines,” Advances in Neural Information Processing Systems 9, pp. 155-161, 1996.
[9]I. T. Jolliffe, “Principal Component Analysis,” Springer, May 1986.
[10]K. Luu, K. Ricanek, T. D. Bui, and C. Y. Suen, “Age estimation using active appearance models and support vector machine regression,” IEEE 3rd International Conference on Biometrics: Theory, Applications, and Systems, pp. 1-5, Sep. 2009.
[11]K. Ricanekand and T. Tesafaye, "MORPH: A Longitudinal Image Database of Normal Adult Age-Progression," IEEE 7th International Conference on Automatic Face and Gesture Recognition, Southampton, pp 341-345, April 2006.
[12]L. G. Farkas, “Anthropometry of the Head and Face,” Raven Press, New York, Jan. 1994.
[13]N. Ramanathan, R. Chellappa, “Modeling Age Progression in Young Faces,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 387-394, Jun. 2006.
[14]N. Hewahi, A. Olwan, N. Tubeel, S. E.-Asar, Z. A.-Sultan, “Age Estimation based on Neural Networks using Face Features,” Journal of Emerging Trends in Computing and Information Sciences, vol. 1, no. 2, Oct. 2010.
[15]S. Yan, H. Wang, T. S. Huang, Q. Yang, and X. Tang, “Ranking with uncertain labels,” IEEE International Conference on Multimedia and Expo, pp. 96-99, Jul. 2007.
[16]Y. Zhang and D.-Y. Yeung, “Multi-Task Warped Gaussian Process for Personalized Age Estimation,” IEEE Conference on Computer Vision and Pattern Recognition, pp. 2622-2629, Jun. 2010.
[17]The MORPH Aging Database [Online], http://www.faceaginggroup.com/.
[18]The FG-NET Aging Database [Online], http://www.fgnet.rsunit.com/.
[19]T. F. Cootes, C. J. Taylor, D. H. Copper, and J. Graham, “Active Shape Model – Their training and Application,” Computer Vision Graphics and Image Understanding, vol. 61, no. 1, pp. 38-59, Jan. 1995.
[20]T. F. Cootes, G. J. Edwards, and C. J. Taylor, “Active Appearance Models,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 23, no. 6, pp. 681-685, Jun. 2001.
[21]X. Geng, Z.-H. Zhou, and Kate Smith-Miles, “Automatic Age Estimation Based on Facial Aging Patterns,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 29, no. 12, pp. 2234-2240, Dec. 2007.
[22]Y. H. Kwon, and N. V. Lobo, “Age Classification from Facial Images,” Computer Vision and Image Understanding Journal, vol. 74, no. 1, pp. 1-21, Apr. 1999.

QR CODE