Image anti-aliasing processing method, image generator training method and device

Abstract

The present invention provides an image anti-aliasing processing method, an image generator training method and device. The image anti-aliasing processing method includes: acquiring an original game image; inputting the original game image to a trained image generator; wherein, the image The generator is a neural network model with a full convolution structure, including: feature extraction network, feature learning network and synthesis network; extract feature information of original game images through feature extraction network; learn and convolute feature information through feature learning network Processing: The feature information and the result of the convolution operation are synthesized through a synthesis network to obtain an anti-aliased game image of the original game image. The image anti-aliasing processing method, image generator training method and device provided by the present invention do not need to perform a large number of operations, so that it can be deployed in equipment that cannot use a large amount of computing resources for anti-aliasing optimization, and is further promoted in computer games.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image anti-aliasing processing method, a training method and device for an image generator. Background technique [0002] Antialiasing methods are commonly used in computer games or other computer programs that generate images to remove image aliasing in rendered images. The anti-aliasing process of an image is actually a process of softening the edges, that is, replacing "precise" with "blur". The existing anti-aliasing algorithms for images are mainly divided into two methods: increasing the sampling rate and image post-processing. [0003] Upsampling (aka supersampling) is a method of spatial antialiasing that works by rendering an image at a much higher resolution than the displayed image, and then using the extra pixels for calculations to reduce it to the desired size. realized. Image post-processing means that the image rendered at the original resol...

AI Technical Summary

Problems solved by technology

[0004] Although the above-mentioned traditional anti-aliasing algorithm has a fast calculation speed, its calculation process depends on powerful computing power, and the number of images in the game is large, and running the game itself requires more hardware resources, so it is difficult to meet this anti-aliasing algorithm. The calculation requirements of the algorithm make it difficult to popularize and apply the anti-aliasing algorithm

Images