Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

real-time texture synthesis

Real-time texture synthesis is the computational process of generating novel visual textures and surface patterns at interactive speeds, typically fast enough to match or exceed standard video frame rates. Unlike offline synthesis techniques that rely on slow, iterative optimization processes, real-time methods use efficient procedural rules, parallelized patch-based sampling, or pre-trained feed-forward neural networks to generate arbitrary-sized texture outputs almost instantaneously. This allows computer graphics systems to create continuous, non-repetitive, high-resolution surface appearances on the fly without the need to store massive image files, supporting applications in video games, virtual environments, dynamic simulations, and live video stylization.

1 item

Precomputed Real-Time Texture Synthesis with Markovian Generative Adversarial Networks

Precomputed Real-Time Texture Synthesis with Markovian Generative Adversarial Networks

Chuan Li, Michael Wand

OrganizationsJohannes Gutenberg University Mainz

Why you should read this

Introduces Markovian Generative Adversarial Networks, a feed-forward approach that eliminates test-time optimization to synthesize textures and stylize videos in real time at speeds hundreds of times faster than previous neural methods.

This paper proposes Markovian Generative Adversarial Networks (MGANs), a method for training generative neural networks for efficient texture synthesis. While deep neural network approaches have recently demonstrated remarkable results in terms of synthesis quality, they still come at considerable computational costs (minutes of run-time for low-res images). Our paper addresses this efficiency issue. Instead of a numerical deconvolution in previous work, we precompute a feed-forward, strided convolutional network that captures the feature statistics of Markovian patches and is able to directly generate outputs of arbitrary dimensions. Such network can directly decode brown noise to realistic texture, or photos to artistic paintings. With adversarial training, we obtain quality comparable to recent neural texture synthesis methods. As no optimization is required any longer at generation time, our run-time performance (0.25M pixel images at 25Hz) surpasses previous neural texture synthesizers by a significant margin (at least 500 times faster). We apply this idea to texture synthesis, style transfer, and video stylization.

Added

2026-09-25