Quick image restoration method of total variation model coupled with gradient fidelity term

Abstract

The invention discloses a quick image restoration method of a total variation model coupled with a gradient fidelity term. The quick image restoration method aims at the TV (Total Variation) model coupled with the gradient fidelity term, and utilizes a splitting algorithm to solve the model in a minimum way to finally realize the quick image restoration. The quick image restoration method comprises the following steps: transforming the model, obtaining an unconstrained transformation model, discretizing the model, and decomposing the model. The quick image restoration method provided by the invention greatly shortens consumed time and realizes the quick image restoration, speed improvement is about five times of a traditional gradient descent, and cost performance is high.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to a fast algorithm of a total variation model coupled with a gradient fidelity item to restore an image. Background technique [0002] Image restoration is a technique to improve image quality, which uses the prior knowledge of degradation factors in the image acquisition model to restore the degraded original image. So far, a large number of image restoration methods have been proposed, which can be roughly divided into four categories: [0003] (1) Inverse filtering and pseudo-inverse filtering methods. Inverse filtering can obtain a better restored image when there is no noise or no zero-crossing point of the modulation transfer function in the image spectrum. Pseudo-inverse filtering methods mainly include generalized inverse method, singular value decomposition pseudo-inverse method and its improved method. Such methods are still sensitive to noise and o...

AI Technical Summary

Problems solved by technology

The disadvantage is that in reality, for some images, the entropy of the original image is often smaller than the entropy of the degraded image, so it is unreasonable to impose the constraint of maximizing the entropy

[0006] (4) The regularization method restoration problem is an inverse problem in mathematics, because the solution of the inverse problem is often ill-posed, that is, the solution does not satisfy the existence, uniqueness and stability

It should be pointed out that the current numerical method of the TV model coupled with the gradient fidelity item is mainly the gradient descent method, which consumes a lot of time and cannot achieve rapid restoration of the image

Images