Image processing model training method, image processing method and related equipment

Abstract

The embodiment of the invention discloses an image processing model training method, an image processing method and related equipment. Wherein the image processing model at least comprises a first network and a second network; the training method comprises the steps of obtaining a sample pair for training, wherein the sample pair comprises a clear image and a blurred image; calling an image processing model to carry out definition recovery processing on the blurred image to obtain a recovered image; updating network parameters of a first network and / or network parameters of a second network inthe image processing model according to the restored image and the clear image; wherein the network parameters of the first network and the network parameters of the second network meet selective sharing conditions, and the selective sharing conditions are used for indicating network parameters for sharing between the first network and the second network and network parameters for indicating mutual independence between the first network and the second network; according to the embodiment of the invention, the image processing model can be better trained, the deblurring performance of the image processing model is optimized, and the model training efficiency is improved.

Description

technical field [0001] The present invention relates to the field of Internet technology, specifically to the field of image processing technology, and in particular to a training method for an image processing model, an image processing method, a training device for an image processing model, an image processing device, a terminal and A computer storage medium. Background technique [0002] Image deblurring is an important research direction in image processing, and its goal is to recover the detail information lost due to blurring in blurred images. With the advancement of research on neural network models, image deblurring methods based on image processing models have achieved better results than traditional methods. The so-called image processing model is a method for deblurring blurred images to obtain clear images neural network model. How to obtain an image processing model with perfect performance through model training is particularly important for the effect of s...

AI Technical Summary

Problems solved by technology

[0003] The inventors found in practice that because the real shooting scene of the blurred image is very complex, which includes multiple factors such as camera motion and object motion in the shooting scene, the existing model training method cannot satisfy the convolution of all motion blur areas. Model assumptions lead to poor image deblurring performance of the trained image processing model; moreover, model training needs to first segment the blurred image and then calculate and then synthesize the blurred image, and the model training efficiency is low

Images