Edge adaptive hybrid norm prior method for blurred image reconstruction

Image deblurring is important for photography and biomédical engineering. There are many existing deblurring methods. However, it is still a challenge to reconstruct images clearly without increasing the effect of noise. In this paper, based on that the probability distributions of gradients vary for different parts of an image, we apply edge adaptation and hybrid norm prior and propose a new deblurring algorithm. Simulations for blurred and noisy images and the biological images acquired from electron microscopes show that, with the proposed algorithm, the reconstructed images can achieve the goals of high clarity and noise robustness at the same time.