An improved correlation filtering based on position prediction

Aiming at the problem that traditional correlation filtering method cannot track fast moving and motion blurred targets very well, a novel correlation filtering based on position prediction algorithm was proposed. Based on the traditional correlation filtering, the prediction based on the target motion vector is introduced to fully consider the target motion information and improve the accuracy of the tracking method. Second, the transform correlation filtering algorithm is used to determine the location of the target in current frame. Third, the correlation filter is used to extract the target area. Finally, the improved algorithm and the traditional algorithm were tracked on four public data sets respectively, and three evaluation criteria were selected to analyze the results. The results show that the proposed algorithm has high accuracy and robustness in tracking fast moving and motion blurred targets.

[1]  Ming-Hsuan Yang,et al.  Incremental Learning for Robust Visual Tracking , 2008, International Journal of Computer Vision.

[2]  Shai Avidan,et al.  Support Vector Tracking , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[3]  Stan Z. Li,et al.  Online Spatio-temporal Structural Context Learning for Visual Tracking , 2012, ECCV.

[4]  Yanxi Liu,et al.  Online Selection of Discriminative Tracking Features , 2005, IEEE Trans. Pattern Anal. Mach. Intell..

[5]  Gang Hua,et al.  Context-Aware Visual Tracking , 2009, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Imran Siddiqi,et al.  Improving handwriting based gender classification using ensemble classifiers , 2017, Expert Syst. Appl..