Improving the Quality of Identifying Personality Attributes from a Face Image by Using a Truncated Learning Sample
DOI:
https://doi.org/10.24160/1993-6982-2018-6-103-109Keywords:
machine learning, local binary patterns, support vector machine, truncated bootstrapping, classification of imagesAbstract
For improving the quality of identifying the attributes of a person, such as sex, race, and age from a face image in the course of sequentially determining these attributes, it is proposed to use the bootstrapping procedure, i.e., learning on "hard" examples, which was previously used for separating the specified object on an image. The bootstrapping procedure involves splitting the training sample into two parts, pre-shaping the decision function based on the first part, classifying the images of the second part using the obtained decision function, and separating all incorrectly classified objects, which are then added to the first part with subsequent retraining. The decision function excessive training effect is reduced by excluding the most difficult to classify examples, which are non-typical representatives of the classified objects, from the training sample formed at the second stage. This approach is called a truncated bootstrapping. The proposed procedure is applied in combination with the support vector machine for carrying out binary classification (determination of sex) and multiple classification (determination of race), and in conjunction with support vector based regression (determination of age). An important feature pertinent to operation of the proposed truncated bootstrapping procedure in the method of sequentially identifying person attributes is that the worst precedents randomly fallen in the training sample are excluded from it: first at the stage of determining the sex, then at the stage of determining the race, and finally at the stage of determining the age. Application of the proposed approach made it possible to improve the quality of producing the decision functions for determining the person attributes when the training sample contains non-typical precedents that fell into it by mistake, and to preserve the accuracy of the classification achieved through sequentially determining the person attributes from the face image.
References
2. Deshmukh D. Gender Recognition from Model’s Face Using SVM Algorithm // Intern. J. Eng. Trends and Techn. 2014. V. 10. No. 1. Pp. 13—18.
3. Demirkus M., Garg K., Guler S. Automated Person Categorization for Video Surveillance Using Soft Biometrics // Proc. Biometrics Techn. for Human Identification. 2010. V. 7667. Pp. 564—576.
4. Zamwar S.C., Ladhake S.A. Human Face Detection and Tracking for Age Rank, Weight and Gender Estimation based on Face Images utilizing Raspberry Pi Processor // Intern. Research J. Eng. and Techn. 2017. V. 4 (2). Pp. 185—188.
5. Shan C. Learning Local Binary Patterns for Gender Classification on Real-World Face Images // Pattern Recognition Lett. 2012. V. 33 (4). Pp. 431—437.
6. Yang Z., Ai H. Demographic Classification with Local Binary Patterns // Proc. IEEE Intern. Conf. Biometrics. 2007. Pp. 464—473.
7. Kazimov T., Mahmudova S. About a Method of Recognition of Race and Ethnicity of Individuals Based on Portrait Photographs // Intelligent Control and Automation. 2014. No. 5. Pp. 120—125.
8. Рыбинцев А.В. Последовательное определение атрибутов личности по изображению лица на основе ранжированных LBP-признаков // Вестник МЭИ. 2017. № 5. С. 121—129.
9. Luu K., Ricanek K., Bui T., Suen C. Age Estimation Using Active Apparance Model and Support Vector Machine Regression // Proc. IEEE 3 rd Intern. Conf. Biometrics: Theory, Appl. and Syst. 2009. Pp. 1—5.
10. Rybintsev A. Age Estimation from a Face Image in a Selected Gender-Race Group based on Ranked Local Binary Patterns // Complex & Intelligent Syst. 2017. V. 3 (2). Pp. 93—104.
11. Maenpaa T. The Local Binary Pattern Approach to Texture Analysis — Extensions and Applications. Oulu University Press, 2003.
12. Рыбинцев А.В. Сокращение размерности пространства LBP-признаков в задачах определения атрибутов личности по изображению лица // Вестник МЭИ. 2016. № 1. С. 33—38.
13. Вьюгин В.В. Математические основы теории машинного обучения и прогнозирования. М.: МЦМНО, 2013.
14. Champagne C., McNairn H., Daneshfar B., Shang J. A Bootstrap Method for Assessing Classification Accuracy and Confidence For Agricul-tural Land Use Mapping in Canada // Intern. J. of Appl. Earth Observation and Geoinformation. 2014. V. 29. Pp. 44—52
15. Felzenszwalb P.F., Girchik R.B., McAllester D., Ramanan D. Object Detection with Discriminatively Trained Part Base Models // IEEE Trans. Pattern Analysis & Machine Intelligence. 2010. V. 32 (9). Pp. 1627—1645.
16. Guo G., Fu Y., Dyer C., Huang T. Image-based Human Age Estimation by Manifold Learning and Locally Adjusted Robust Regression // IEEE Trans. Image Proc. 2008. V. 17 (7). Pp. 1178—1188.
17. Chang K-Y., Chen C-S. A Learning Framework for Age Rank Estimation based on Face Images with Scattering Transform // IEEE Trans. Image Proc. 2015. V. 24 (3). Pp. 785—798.
18. Liu H., Lu J., Feng J., Zhou J. Group-aware Deep Feature Learning for Facial Age Estimation // Pattern Recognition. 2017. V. 66. Pp. 82—94.
---
Для цитирования: Рыбинцев А.В. Повышение качества определения атрибутов личности по изображению лица путем усечения обучающей выборки // Вестник МЭИ. 2018. № 6. С. 103—109. DOI: 10.24160/1993-6982-2018-6-103-109.
#
1. Moghaddam B., Yang M. Learning Gender with SuPp.rt Faces. IEEE Trans. and Pattern Analysis and Machine Intelligence. 2002;24 (5):707—711.
2. Deshmukh D. Gender Recognition from Model’s Face Using SVM Algorithm. Intern. J. Eng. Trends and Techn. 2014;10;1:13—18.
3. Demirkus M., Garg K., Guler S. Automated Person Categorization for Video Surveillance Using Soft Biometrics. Proc. Biometrics Techn. for Human Identification. 2010;7667:564—576.
4. Zamwar S.C., Ladhake S.A. Human Face Detection and Tracking for Age Rank, Weight and Gender Estimation based on Face Images utilizing Raspberry Pi Processor. Intern. Research J. Eng. and Techn. 2017;4 (2):185—188.
5. Shan C. Learning Local Binary Patterns for Gender Classification on Real-World Face Images. Pattern Recognition Lett. 2012;33 (4):431—437.
6. Yang Z., Ai H. Demographic Classification with Local Binary Patterns. Proc. IEEE Intern. Conf. Biometrics. 2007:464—473.
7. Kazimov T., Mahmudova S. About a Method of Recognition of Race and Ethnicity of Individuals Based on Portrait Photographs. Intelligent Control and Automation. 2014;5:120—125.
8. Rybintsev A.V. Posledovatel'noe Opredelenie Atributov Lichnosti po Izobrazheniyu Litsa na Osnove Ranzhirovannyh LBP-priznakov. Vestnik MPEI. 2017;5:121—129. (in Russian).
9. Luu K., Ricanek K., Bui T., Suen C. Age Estimation Using Active Apparance Model and Support Vector Machine Regression. Proc. IEEE 3 rd Intern. Conf. Biometrics: Theory, Appl. and Syst. 2009:1—5.
10. Rybintsev A. Age Estimation from a Face Image in a Selected Gender-Race Group based on Ranked Local Binary Patterns. Complex & Intelligent Syst. 2017;3 (2):93—104.
11. Maenpaa T. The Local Binary Pattern Approach to Texture Analysis — Extensions and Applications. Oulu University Press, 2003.
12. Rybintsev A.V. Sokrashchenie Razmernosti Prostranstva LBP-priznakov v Zadachah Opredeleniya Atributov Lichnosti po Izobrazheniyu Litsa. Vestnik MPEI. 2016;1:33—38. (in Russian).
13. V'yugin V.V. Matematicheskie Osnovy Teorii Mashinnogo Obucheniya i Prognozirovaniya. M.: MTSMNO, 2013. (in Russian).
14. Champagne C., McNairn H., Daneshfar B., Shang J. A Bootstrap Method for Assessing Classification Accuracy and Confidence For Agricul-tural Land Use Mapping in Canada. Intern. J. of Appl. Earth Observation and Geoinformation. 2014;29:44—52
15. Felzenszwalb P.F., Girchik R.B., McAllester D., Ramanan D. Object Detection with Discriminatively Trained Part Base Models. IEEE Trans. Pattern Analysis & Machine Intelligence. 2010;32 (9):1627—1645.
16. Guo G., Fu Y., Dyer C., Huang T. Image-based Human Age Estimation by Manifold Learning and Locally Adjusted Robust Regression. IEEE Trans. Image Proc. 2008;17 (7):1178—1188.
17. Chang K-Y., Chen C-S. A Learning Framework for Age Rank Estimation based on Face Images with Scattering Transform. IEEE Trans. Image Proc. 2015;24 (3):785—798.
18. Liu H., Lu J., Feng J., Zhou J. Group-aware Deep Feature Learning for Facial Age Estimation. Pattern Recognition. 2017;66:82—94.
---
For citation: Rybintsev A.V. Improving the Quality of Identifying Personality Attributes from a Face Image by Using a Truncated Learning Sample. MPEI Vestnik. 2018;6:103—109. (in Russian). DOI: 10.24160/1993-6982-2018-6-103-109.
