Image denoising method based on adversarial generative network

Abstract

The invention discloses an image denoising method based on an adversarial generative network, and the method is characterized in that the method comprises the following steps: step1, carrying out thepreparation of a data set; step 2, constructing a generative network: taking the input of the generative network as a noise image, extracting noise characteristics, and performing image denoising processing; step 3, constructing a discrimination network: determining that the input of the discrimination network is a clear image without noise and an image generated by a generation network, and thenperforming discrimination; step 4, training the generative adversarial network by using a noise-free clear image and a noise image according to the idea of the generative adversarial network to obtaina trained image denoising adversarial generation network, and storing the trained parameters; wherein the generative adversarial network comprises a generative network and a discriminant network; andstep 5, inputting the noise image into the trained generation network to obtain a denoised image. According to the method provided by the invention, manual feature extraction is not needed, manual intervention is reduced, and the quality of the denoised image is ensured.

Description

technical field [0001] The invention relates to the fields of computer graphics and artificial intelligence, in particular to an image denoising method based on an adversarial generation network. Background technique [0002] During the imaging and transmission process, the image is affected by photoelectric conversion, digital-to-analog conversion, environmental noise, channel noise and other issues, which will cause the quality of the image to deteriorate. Image noise will affect the accurate acquisition of image information, which will have a serious impact on image target detection, recognition, image segmentation and other research. Especially with the continuous development of society, high-quality images are required for further processing in various fields, such as: security, medical images, industrial inspection, etc. The quality of images has a decisive effect on them. Therefore, image denoising is of great significance. [0003] There have been many attempts at ...

AI Technical Summary

Problems solved by technology

However, there are still problems such as insufficient restoration of image details and texture features, and it is only applicable to specific noises, and cannot achieve blind denoising of image noise.

Images