Merging and Arbitration Strategy for Robust Eye Location
The difficulties of eye location are mainly caused by the variations of intrinsic eye pattern characteristics from people to people, scale, pose, glasses frame, illumination, etc. To prevail from these problems, this paper addresses a novel and precise robust eye location method. It employs appearance based Bayesian framework to relive the effect of uneven illumination. The appearance of eye patterns is represented by 2D Haar wavelet. It also employs a sophisticated merging and arbitration strategy in order to manage the variations in geometrical characteristics of ambient eye regions due to glasses frames, eye brows, and so on. The located eye candidates are merged or eliminated according to the merging rule. If the merged regions are more than one, we apply the arbitration strategy. The arbitration strategy is based on a minimizing energy function by probabilistic forces and image forces that pull it toward eyes. The experimental results show that the proposed approach can achieve superior performance using various data sets to previously proposed methods.
KeywordsMahalanobis Distance Bayesian Classifier Probabilistic Force Merging Strategy Uneven Illumination
Unable to display preview. Download preview PDF.
- 4.Rowley, H.A., Baluja, S., Kanade, T.: Neural Network-Based Face Detection. IEEE Trans. Pattern Analysis and Machine Intelligence 20(1) (1998)Google Scholar
- 5.Zhou, H., Geng, X.: Projection functions for eye detection. Pattern Recognition (in press, 2004)Google Scholar
- 6.Ma, Y., Ding, X., Wang, Z., Wang, N.: Robust precise eye location under probabilistic framework. In: IEEE International Conference on Automatic Face and Gesture Recognition (2004)Google Scholar
- 7.Ichihashi, H., Honda, K., Wakami, N.: Robust PCA with Intra-sample Outlier Process Based on Fuzzy Mahalanobis Distances and Noise Clustering. In: IEEE International Conference on Fuzzy Systems (2005)Google Scholar
- 8.Watabe, A., Komiya, K., Usuki, J., Suzuki, K., Ikeda, H.: Effective Designation of Specific Shots on Video Service System Utilizing Mahalanobis Distance. IEEE Transactions on Consumer Electronics 51(1), (February 2005)Google Scholar
- 9.Zuo, F., Real-time face recognition for smart home applications. In: Consumer Electronics, 2005. ICCE. 2005 Digest of Technical Papers. International Conference, January 8-12, 2005, pp. 35–36 (2005) Google Scholar
- 11.Huang, J., Shao, X.H., Wechsler, H.: Pose Discrimination and Eye Detection Using Support Vector Machines. In: Proceeding of NATO-ASI on Face Recognition: From Theory to Applications (1998)Google Scholar
- 12.Viola, P., Jones, M.: Rapid object detection using a Boosted cascade of simple features. In: Proc. of IEEE Conf. on CVPR, pp. 511–518 (2001)Google Scholar
- 13.Schneiderman, H., Kanade, T.: A statistical model for 3D object detection applied to faces and cars. In: Proc. of IEEE Conf. on CVPR (2000)Google Scholar
- 14.Kawaguchi, T., Hikada, D., Rizon, M.: Detection of the eyes from human faces by hough transform and separability filter. In: Proc. of ICIP, pp. 49–52 (2000)Google Scholar
- 17.Lucey, S., Sridharan, S., Chandran, V.: Improved facial feature detection for AVSP via unsupervised clustering and dicriminant analysis. EURASIP Journal on Applied Signal Processing 3, 264–275 (2003)Google Scholar