Artificial intelligence usage for the formation of graphic images

Abstract

In the work, it is analyzed the main aspects of artificial intelligence usage for the formation of graphic images. The main stages of the graphics pipeline for image formation and the prospects of applying artificial intelligence for their optimization are considered. Generative adversarial networks, diffusion models, transformers, ensemble learning methods, which are used for image generation, are characterized. Direct image generation as the alternative to the image formation using graphics editing tools is examined. The ways of optimizing the stages of 3D-model formation, texture generation, texture mapping, 3D-model`s surface shading and formed image post-processing are analyzed in detail.