A novel and blur-invariant local feature image matching

摘要

The lack of robustness to blur (especially motion blur) is one of the biggest problem in the existing local feature schemes. In this paper, we present a novel Local feature scheme to solve this problem. Our proposed method is good at image matching between (motion and Gaussian) blurred images and non-blurred images. Experimental results show that the proposed method outperforms state of the art methods for blurred image matching.