Face Detection Using Multi-Feature

In this paper, we propose a novel method called Multi-Feature Soft Cascade Learning for improving the performance of face detection. The main contribution of this paper consists of the following two aspects. The first is the use of Multi-Feature in AdaBoost, resulting in a more stable boost classifier with fewer features compared with using only single features as well as the improvement of the detection performance. The second is the new soft cascade algorithm for the Multi-Feature training, which works together with the Multi-Feature selection criterion. Experiment results show the improvement by using Multi-Feature compared with single feature. We also find that the candidate feature set is another important factor to improve the face detection performance.

[1]  Mingjing Li,et al.  Robust multipose face detection in images , 2004, IEEE Transactions on Circuits and Systems for Video Technology.

[2]  Paul A. Viola,et al.  Rapid object detection using a boosted cascade of simple features , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[3]  Rong Xiao,et al.  Boosting chain learning for object detection , 2003, Proceedings Ninth IEEE International Conference on Computer Vision.

[4]  Hanqing Lu,et al.  Face detection using improved LBP under Bayesian framework , 2004, Third International Conference on Image and Graphics (ICIG'04).

[5]  Jonathan Brandt,et al.  Robust object detection via soft cascade , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[6]  Yair Weiss,et al.  Learning object detection from a small number of examples: the importance of good features , 2004, CVPR 2004.

[7]  Henry Schneiderman,et al.  Learning a restricted Bayesian network for object detection , 2004, CVPR 2004.

[8]  Akinobu Shimizu,et al.  Classification-based face detection using Gabor filter features , 2004, Sixth IEEE International Conference on Automatic Face and Gesture Recognition, 2004. Proceedings..