Image deblurring method based on generative adversarial network

Abstract

The invention discloses an image deblurring method based on a generative adversarial network. The method mainly solves the problems that in the prior art, when a blurring kernel needs to be estimatedand a clear image needs to be solved through deconvolution iteration, the operation speed is low, and the reconstruction effect is poor. The implementation scheme is as follows: 1) selecting an experimental data set, and determining a training data set and a test data set related to the experimental data set; 2) respectively constructing a generation network with a 15-layer structure and an adversarial network with a 6-layer structure; 3) constructing a joint loss function according to the adversarial loss, the pixel loss and the feature loss; 4) performing interactive training on the generative adversarial network and the adversarial network through a joint loss function to obtain a generative adversarial network model; and 5) inputting the test sample into the generative adversarial network model to obtain a deblurred clear image The method has the advantages of no need of estimating a blurring kernel, high deblurring speed and good deblurring effect, and can be used for deblurring processing of blurred images shot due to camera jittering.

Description

Technical field [0001] The present invention belongs to the technical field of image processing, and further relates to an image deblurring method, which can be used for deblurring processing of blurred images shot due to camera shake. Background technique [0002] In the process of digital image generation, processing, transmission, and storage, the overall quality of the image will deteriorate due to the influence of the camera and the unpredictable factors outside, and the important information of the image will be lost, which will bring many negative effects on subsequent processing. In order to meet the requirements of image applications and image vision for high-quality pictures, how to recover clear images from low-quality images has always been a long-term concern in the field of digital imaging. [0003] In the early stage of image restoration, researchers mainly studied non-blind deblurring techniques, and proposed methods such as Wiener filtering, LR iterative deblurring...

AI Technical Summary

Problems solved by technology

Because these two methods still use the idea of ​​non-blind deblurring algorithm, after obtaining the estimated blur kernel through CNN, they use the traditional method to deconvolute to obtain a clear image, which leads to slow running speed and the reconstruction effect depends heavily on the blur kernel. Predict the outcome, and cannot act on multiple fuzzy functions

Images