Hybrid Order l0-Regularized Blur Kernel Estimation Model for Image Blind Deblurring

• Published in: Advances in Neural Networks - ISNN 2017 Vol. 10262; pp. 239 - 247
• Main Authors: Li, Weihong, Chen, Yangqing, Chen, Rui, Gong, Weiguo, Zhao, Bingxin
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2017; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Adaptive adjustment factor; Blur kernel estimation; Half-quadratic splitting technique; Hybrid order; Image blind deblurring; regularized
• ISBN: 9783319590806; 3319590804
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-319-59081-3_29

Most of blur kernel estimation models may fail when the blurred image contains some complex structures or is contaminated by large blur. In this paper, we propose a hybrid order l0-regularized blur kernel estimation model for solving the problem. Firstly, we regularize the latent image in a hybrid order case involving both first-order and second-order regularization term, in which l0 sparse prior is introduced. Secondly, we introduce an improved adaptive adjustment factor into the model for removing detrimental structures and obtaining more useful information. Finally, we develop an efficient optimization algorithm based on the half-quadratic splitting technique to obtain an accurate blur kernel. Extensive experiments results on both synthetic and some challenged real-life images show that proposed model can estimate a more accurate blur kernel and can effectively recover the latent image when it contains complex structures or is contaminated by large blur.