臺灣博碩士論文加值系統

這篇論文提出一個藉由人臉皮膚色的資訊以加強人臉偵測準確度的方法。大部份使用boosting的演算法使用了間單的Haar features，而且只使用了一張圖灰階的資訊(intensity)。雖然這些features很有效，但是相片中的彩色資訊往往被忽略。只使用灰階資訊的features容易被某些不同的光原影響，例如從不同方向的光原會直接造成不同區域上的陰影。利用了皮膚色資訊(skin color)，此類的光原影響可以被降低，因為與灰階資訊比較起來，膚色資訊比較不會直接受亮度影響。在我們的演算法裡，我們先將一張圖轉成灰階圖(intensity map)以及膚色圖(skin color map)。取得膚色圖的方法是使用了一個Multi-Layer Perceptron的演算法。而我們的人臉偵測系統同時使用了在灰階圖以及膚色圖下取得的Haar features，並用Adaptive Boosting的方法以結合此兩種features。我們將我們的演算法套用在不同角度的人臉偵測，並在實驗中，與只使用灰階圖的features相比，我們的演算法得到較好的效果。

This thesis describes a face detection system that improves the performance of the boost-based algorithms by introducing a novel set of features based on skin color. Most boost-based algorithms use a boosted cascade of simple Haar features on intensity values. Though effective, these features completely ignore the color information. Thus, the effectiveness of these features inevitably suffers from the variance in lighting conditions, such as the direction of lighting and as well as the shadowing effects. With the incorporation of skin color information, the difficulty of such illumination variations may be reduced in that the skin color information is less sensitive to changes in brightness. As for our pattern classifier, in addition to the intensity map, the input image is converted into a skin color map by calculating the regression values using the Multi-Layer Perceptron (MLP) algorithm. Weak classifiers are then constructed based on Haar features of both the intensity map and the skin color map. We apply our skin-enhanced algorithm to detect faces of various poses and we show that our algorithm achieves better performance than the method based on only intensity Haar features.

Chapter 1 Introduction 1
1.1 Face Detection 1
1.2 Motivation 2
1.3 System Overview 4
1.4 Thesis Organization 5
Chapter 2 Related Work 7
2.1. Skin Color Classification 7
2.2. Face Detection Algorithms 9
2.3 Multiview Face Detection 11
Chapter 3 Skin Color Classifier 15
3.1 Multi-Layer Perceptron Skin Color Classifier 15
3.2 Training and Experiment 17
Chapter 4 AdaBoost Face Detector 21
4.1 AdaBoost Learning 21
4.2 Skin Haar features 24
4.3 Post-processing 26
4.4 Multiview Face Detection 26
Chapter 5 Experiments 29
5.1 Evaluation of Frontal Face Detector 29
5.2 Evaluation of Multiview Face Detectors 32
Chapter 6 Conclusion and Future Work 39
6.1 Conclusion 39
6.2 Future Work 40
Reference 43

[1]S. B. H. A. Rowley and T. Kanade. “Neural network-based face detection”. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(1):23-28, 1998.
[2]H. Rowley, S. Buluja, and T. Kanade. “Rotation invariant neural network-based Face Detection”. In Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1998.
[3]S. Z. Li and A. K. Jain. Handbook of Face Recognition. Springer, 2005.
[4]S. Z. Li and Z. Zhang. “FloatBoost learning and statistical face detection”. IEEE Transactions on Pattern Analysis and Machine Intelligence, 26(9):1112-1123, 2004.
[5]R. Lienhart, A. Kuranov, and V. Pisarevsky. “Empirical analysis of detection cascades of boosted classifiers for rapid object detection”. MRL Technical Report, Intel Labs, 2002.
[6]R. Lienhart and J. Maydt. “An Extended Set of Haar-like Features for Rapid Object Detection”. In Proceedings of IEEE, ICIP, pages 900-903, 2002.
[7]S. L. Phung, A. Bouzerdoum, and D. Chai. “Skin Segmentation Using Color Pixel Classification: Analysis and Comparison”. IEEE Transaction on Pattern Analysis and Machine Intelligence, 27(1):148-154, 2005.
[8]S. L. Phung, D. Chai, and A. Bouzerdoum. “A universal and robust human skin color model using neural networks”. In Proceedings of INNS-IEEE International Joint Conference on Neural Networks, pages 2844-2849, 2001.
[9]K.-K. Sung and T. Poggio. “Example-based Learning for View-Based Human Face Detection”. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(1):39-51, 1998.
[10]M. A. Turk and A. P. Pentland. “Eigenfaces for recognition”. Journal of Cognitive Neuroscience, 3(1):71-86, 1991.
[11]P. Viola and M. J. Jones. “Rapic Object Detection Using a Boosted Cascade of Simple Features”. In Proceedings of IEEE CVPR 2001, pages 609-615, December 2001.
[12]M. Jones and P. Viola, “Fast multi-view face detection”. MERL-TR2003-96, July 2003.
[13]R. Xiao, M. Li, and H. Zhang. “Robust multi-pose face detection in images”. IEEE Transactions on Circuits and Systems for Video Technology, 14(1):31-41, 2004.
[14]M.-H. Yang, D. Kriegman, and N. Ahuja. “Detecting faces in images: a survey”. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24(1):34-58, 2002.
[15]B. Menser and F. Muller. “Face detection in color images using principal components analysis”. Image Processing and Its Applications, 1999.
[16]R.-L. Hsu, M. Abdel-Mottaleb, and A. K. Jain. “Face detection in color images”. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24(5):696-706, 2002.
[17]C. Huang, H. Ai, Y. Li, and S. Lao. “High-performance rotation invariant multiview face detection”. IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 29, No. 4, April 2007.
[18]C. Huang, B. Wu, H. Ai, and S. Lao. “Omni-directional face detection based on Real AdaBoost”. ICIP: 593-596, 2004.
[19]Y.-Y. Lin and T.-L. Liu. “Robust face detection with multi-class boosting”. In Proceedings of the 2005 IEEE CVPR, 2005.
[20]H. Schneiderman and T. Kanade. “A statistical method for 3D object detection applied to faces and cars”. Proceedings IEEE Conference on Computer Vision and Pattern Recognition, vol.1, pp. 746-751, 2000.
[21]Open Computer Vision Library. http://sourceforge.net/projects/opencvlibrary/