Robust Tracking for Motion Blur Based on Correlation Filter

Correlation filter methods show good performance in robustness, accuracy and speed in visual tracking. In this paper, we propose a new algorithm based on Kernelized Correlation Filter (KCF) to effectively improve performance on motion blur. We adopt a simple and effective image deblurring method based on dark channel prior, combined with correlation filter to form a new Deblurring Correlation Filter (DECF) tracking algorithm. Extensive experimental results on OTB2015 show that our tracker demonstrates the excellent accuracy on tracking with motion blur.