A gradient histogram preservation based texture enhanced model for image deblurring

Image blurring attenuate crucial textures and thus always result in a pretty dismal visual experience. Unfortunately, image blurring is difficult to avoid during the image acquisition, hence, lots of recent research focus on how to preserve subtle textures while suppress visual artifacts during image deblurring. Among all of the existing image deblurring methods, image priors, such as non-local priors of image gradient, play an important role. Although using prior knowledge improves noise and ringing artifact removal, fine textures always are attenuated as noise or ringing artifact. In order to solve this problem, we introduce a gradient histogram preservation (GHP) based deblurring model. Combining the GHP model with the non-local sparse prior, we impose both a global constraint and a non-local sparse constraint, and are capable of synthesizing rich textures. We introduce an alternative image deblurring scheme, the problem can be separated into several sub-problems and alternatively solved. Our deblurring method constraints the gradient distribution of restored images to approaching the gradient distribution of the latent sharp image. Furthermore, the experimental results on both uniform and non-uniform blurry images substantiate that the proposed deblurring method performs superior to most of the state-of-the-arts.