AdaBoost is a well developed method for classification, and has been applied to face detection. Given a large set of training data, AdaBoost-based face detector combines many weak classifiers, each of which is targeted at a set of features with different characteristics, and generates a strong classifier able to detect faces with large variations on the appearances. However, when the patterns embedded in the positive and negative training data conflict to each other, the training can often become time-consuming, and in some cases, be extremely difficult to converge. This research proposes a cascade architecture for AdaBoost, known as Cascade AdaBoost, which decomposes the training data into clusters with similar features. A component face detector is obtained using AdaBoost classification on each cluster of data, and the combination of all component face detectors contributes to the overall face detector. This research also studies the impacts of features with gray-scale pixel values and pixels with skin colors. Experiments on PIE and FRGC databases reveal that the proposed Cascade AdaBoost face detector can converge much faster than the conventional AdaBoost method, and that the skin color can substantially improve the accuracy and speed of face detection.

AdaBoost is a well developed method for classification, and has been applied to face detection. Given a large set of training data, AdaBoost-based face detector combines many weak classifiers, each of which is targeted at a set of features with different characteristics, and generates a strong classifier able to detect faces with large variations on the appearances. However, when the patterns embedded in the positive and negative training data conflict to each other, the training can often become time-consuming, and in some cases, be extremely difficult to converge. This research proposes a cascade architecture for AdaBoost, known as Cascade AdaBoost, which decomposes the training data into clusters with similar features. A component face detector is obtained using AdaBoost classification on each cluster of data, and the combination of all component face detectors contributes to the overall face detector. This research also studies the impacts of features with gray-scale pixel values and pixels with skin colors. Experiments on PIE and FRGC databases reveal that the proposed Cascade AdaBoost face detector can converge much faster than the conventional AdaBoost method, and that the skin color can substantially improve the accuracy and speed of face detection.

1. M. Yang, D. J. Kriegman, N. Ahujo, ”Detecting Faces in Images: A survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, no 1, January 2002.
2. L. G. Valiant, “A theory of the learnable, ” Communications of the ACM, vol. 27, no. 11, pp. 1134–1142, November 1984.
3. M. J. Kearns and U. Z. Vazirani, “An Introduction to Computational Learning Theory,” MIT Press, 1994.
4. M. Kearns and L. G. Valiant, ”Learning Boolean formulae or finite automata is as hard as factoring,” Technical Report TR-14-88, Harvard University Aiken Computation Laboratory, August 1988.
5. M. Kearns and L. G. Valiant., “Cryptographic limitations on learning Boolean formulae and finite automata,” Journal of the Association for Computing Machinery, vol. 41, no. 1, pp. 67–95, January 1994.
6. R. E. Schapire., “The strength of weak learn ability, ” Machine Learning, vol. 5, no. 2, pp.197–227, 1990.
7. Y. Freund., “Boosting a weak learning algorithm by majority, ” Information and Computation, vol. 121, no. 2, pp. 256–285, 1995.
8. H. Drucker, R. Schapire, and P. Simard., “Boosting performance in neural networks,” International Journal of Pattern Recognition and Artificial Intelligence, vol. 7, no. 4, pp. 705-719, 1993.
9. P. Deng and M. Pei, ”Multi view Face Detection Based on AdaBoost and skin color,” IEEE First International Conference on Intelligent Networks and Intelligent Systems (ICINIS), pp. 457-460, 2008.
10. D. Huang, W. Hu, and M. Hsu, “Adaptive Skin Color Model Switching for Face Tracking under Varying Illumination”, IEEE Fourth International Conference on Innovative Computing Information and Control (ICICIC), pp. 326-329, 2009.
11. S. A. Inalou and S. Kasaei, "AdaBoost-Based Face Detection in Color Images with Low False Alarm," Second International Conference on Computer Modeling and Simulation, vol. 2, pp. 107-111, 2010.
12. R. Hsu. and M. Abdel-Mottaleb, ”Face Detection in Color Images,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, no. 5, pp. 696-706, 2002.
13. K. Ichikawa, T. Mita, et all, “Component-based Face Detection methods for Various Types of Occluded Faces,” 3rd International Symposium on Communications, Control, and Signal Processing (ISCCSP), 2008.
14. L. Goldman, U. J. Monich, and Sikora T., “Component and Their Topology for Robust Face Detection in the presence of partial occlusion,” IEEE Transactions on Information Forensics and Security, vol. 2, no. 3, pp. 559-569, 2007.
15. J. Chen et all, “Modification of the AdaBoost-based Detector for partially Occluded Face,” Proceedings - International Conference on Pattern Recognition 2, pp. 516-519, 2006.
16. P. Viola and M. J. Jones, “Robust Real Time Face Detection,” International Journal of Computer Vision, vol. 57, no. 2, pp. 137–154, 2004.
17. P. Viola, and M. Jones, “Robust Real Time Face Detection,” International Journal of Computer Vision, vol.57, no. 2, pp. 137–154, 2004.
18. Y. Shi, F. Zhao, and Z. Zhang, “Hardware Implementation of AdaBoost Algorithm and verification,” International Conference on Advanced Information Networking and Applications (AINA), pp. 343-346, 2008.
19. V. Galtier, et all, ”AdaBoost Parallelization on PC clusters with virtual shared memory for fast feature selection”, International Conference on Signal Processing and Communication (ICSPC), pp. 165-168, IEEE 2007.
20. Y. Freund and R. E. Schapire, “Game theory, on-line prediction and boosting,” Proceedings of the Ninth Annual Conference on Computational Learning Theory, pp. 325–332, 1996.
21. J. Friedman, T. Hastie, and R. Tibshirani, “Additive logistic regression: A statistical view of boosting,” Annals of Statistics vol. 28, no. 2, pp. 337-407, 2000.
22. A. Vezhnevets and V. Vezhnevets, “Modest AdaBoost’ – Teaching AdaBoost to Generalize better”, Moscow State University, retrieved at June 24, 2010, retrieved at June 17th, 2010, from http://graphics.cs.msu.ru/en/science/research/machinelearning/adaboosttoolbox.
23. G. Leshem, Software AdaBoost, retrieved at June 24th, 2010, from http://shum.huji.ac.il/~gleshem/.
24. Graphic and media Lab. CMC department, Moscow State University, retrieved at June 17th, 2010, from http://graphics.cs.msu.ru/en/science/research/machinelearning/adaboosttoolbox.
25. C. Mertayak, Matlab Central, retrieved at June 24th, 2010, from http://www.mathworks.com/matlabcentral/fileexchange /21317-adaboost.
26. R. Lienhart, and J. Maydt, “An Extended Set of Haar-like Features for Rapid Object Detection,” IEEE International Conference on Image Processing 1, pp. I/900-I/903, 2002.
27. D. Dervinis, “Head Orientation Estimation using Characteristic Points of Face,” The 3rd Conference on Communication by Gaze Interaction – COGAIN, pp. 45-48, 2007.
28. Adelson et all, “Pyramid methods in image processing,” RCA engineer 29 (6), pp. 33-41, 1984.
29. YCbCr, wikipedia.org, retrieved at June 25th, 2010, from http://en.wikipedia.org/wiki/YCbCr.
30. C. A Poynton, “A Technical Introduction to Digital Video,” John Wiley & Sons, Inc., pp. 17, 1996,
31. Rec. ITU-R BT.601-5, “Studio Encoding Parameters of Digital Television for Standard 4:3 and Wide-screen 16:9 Aspect Ratios,” (1982-1986-1990-1992-1994-1995), Section 3.5.
32. C. Garcia and G. Tziritas, “Face Detection Using Quantized Skin Color Regions Merging and Wavelet Packet Analysis,” IEEE Transactions on multimedia, vol. 1, no. 3, 1999.
33. T. Sim, S. Baker, and M. Bsat, “The CMU pose, illumination, and expression database,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 25, no. 12, pp. 46-51, 2003.
34. P.J. Philips, and P.J. Flynn et all, “Overview of the face recognition grand challenge,” IEEE conf. computer vision and pattern recognition, vol. 1, pp. 947-954, 2005.