Blurred/Non-Blurred Image Alignment using Sparseness Prior

Aligning a pair of blurred and non-blurred images is a prerequisite for many image and video restoration and graphics applications. The traditional alignment methods such as direct and feature-based approaches cannot be used due to the presence of motion blur in one image of the pair. In this paper, we present an effective and accurate alignment approach for a blurred/non-blurredimage pair. We exploit a statistical characteristic of the real blur kernel - the marginal distribution of kernel value is sparse. Using this sparseness prior, we can search the best alignment which produces the sparsest blur kernel. The search is carried out in scale space with a coarse-to-fine strategy for efficiency. Finally, we demonstrate the effectiveness of our algorithm for image deblurring, video restoration, and image matting.

• Type: Conference
• Year: 2007
• Where: ICCV
• Authors: Lu Yuan, Jian Sun, Long Quan, Heung-Yeung Shum
• Keywords: Accurate Alignment Approach, Computer Vision, ICCV 2007, Motion Blur, Real Blur Kernel, Sparsest Blur Kernel, Traditional Alignment Methods