Hybrid Feature and Adaptive Particle Filter for Robust Object Tracking

Authors: Xinyue Zhao, Yutaka Satoh, Hidenori Takauji, Shun'ichi Kaneko

Abstract:

A hybrid feature based adaptive particle filter algorithm is presented for object tracking in real scenarios with static camera. The hybrid feature is combined by two effective features: the Grayscale Arranging Pairs (GAP) feature and the color histogram feature. The GAP feature has high discriminative ability even under conditions of severe illumination variation and dynamic background elements, while the color histogram feature has high reliability to identify the detected objects. The combination of two features covers the shortage of single feature. Furthermore, we adopt an updating target model so that some external problems such as visual angles can be overcame well. An automatic initialization algorithm is introduced which provides precise initial positions of objects. The experimental results show the good performance of the proposed method.

Keywords: Hybrid feature, adaptive Particle Filter, robust Object Tracking, Grayscale Arranging Pairs (GAP) feature.

Digital Object Identifier (DOI): doi.org/10.5281/zenodo.1057687

References:

[1] N. Dalal and B. Triggs. Histograms of oriented gradients for human detection. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), pages 886-893, 2005.
[2] K. Nummiaro, E. Koller-Meier, and L. Van-Gool. An adaptive color-based particle filter. Image and Vision Computing, 21:99-110, 2003.
[3] P. P'erez, C. Hue, J. Vermaak, and M. Gangnet. Color-based probabilistic tracking. In European Conference on Computer Vision (ECCV), pages 661-675, 2002.
[4] J. Wang, X. Chen, and W. Gao. Online selecting discriminative tracking features using particle filter. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), pages 1037-1042,2005.
[5] Xinyue Zhao, Yutaka Satoh, Hidenori Takauji, Shun-ichi Kaneko, Kenji Iwata, and Ryushi Ozaki. Robust moving object detection based on a statistical model. In The Sixteenth Korea-Japan Joint Workshop on Frontiers of Computer Vision, pages 283-288, 2010.
[6] Xinyue Zhao, Yutaka Satoh, Hidenori Takauji, Shun-ichi Kaneko, Kenji Iwata, and Ryushi Ozaki. Object detection based on a robust and accurate statistical multi-point-pair model. Pattern Recognition, 44(6):1296-1311, 2011.