Spatial-scale-based blur kernel estimation for blind motion deblurring

Shu Tang, Xianzhong Xie, Xiao Luan, Ming Xia, Peisong Liu

Abstract: Maximum a posteriori (MAP)-based single-image blind motion deblurring methods are extensively studied in the past years, and have achieved great progress. However, because of imperfect salient edges selection, most state-of-the-art methods still cannot estimate the blur kernel (BK) accurately, especially in large motion blur cases. In this paper, we propose a novel spatial-scale-based approach to estimate an accurate BK from a single motion blurred image by combining the spatial scale and L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> norm. Furthermore, we propose an efficient optimization strategy which can solve the proposed model efficiently. Extensive experiments compared with state-of-the-art blind motion deblurring methods demonstrate the effectiveness of our method.