Progressive Recovery of Image through Hybrid Graph Laplacian Regularization

The problem of image restoration has a long and well-travelled history. Image restoration is still a valid challenge. The two main limitations in image accuracy are noise and blur. Image restoration includes removing noise from the image and removing the blur from the image. This paper proposes a unified framework for performing image denoising and deblurring. The restoration task is performed progressively and the task of restoration executed in a repeated manner. The number of repetition is based on the noise or blur level, and then the task of restoration is performed. In this way it recovers more and more image details and edges. We test our algorithm based on psnr value and it shows a higher performance than state-of-the-art algorithms.