2 min read

A team of researchers have discovered a new tool that can assist designers and animators to create more realistic virtual textures. The tool makes uses of a deep learning technique, called the Generative Adversarial Networks (GANs). GANs train a neural network to learn to expand small textures into larger ones that bear a resemblance to the original sample.

Texture synthesis has always remained a challenging job for designers. The design of accurate real-world textures such as water ripples in a river, concrete walls, or patterns of leaves is highly intricate and prone to errors. Current techniques for texture creation are tedious and time-consuming. However, the works of Yang Zhou et al, have made the texture synthesis process simplistic for texture artists in designing video games, virtual reality, and animation.

Their method uses a generator to generate a texture, usually larger in size than the input, that closely resembles the visual characteristics of the sample input. The visual similarity between the newly created texture and the sample input is assessed using a discriminative network (discriminator). As typical of GANs, the discriminator is trained in parallel to the generator to distinguish between the actual and the desired output.

The researchers tested their method on complex examples of peacock feathers and tree trunk ripples, which are seemingly endless in their repetitive patterns. The results are realistic designs produced in high-resolution, efficiently, and at a much larger scale.

The team also intends to train a “universal” model on a large-scale texture dataset, as well as increase user control as part of their future work.

Zhou and his collaborators will present their work at SIGGRAPH 2018, to be held on  12-16 August in Vancouver, British Columbia. This annual gathering showcases the works of professionals, and academicians practicing in CG, Animation, VR, Games, Digital Art, Mixed Reality and Emerging Technologies.

Read Next

Nvidia and AI researchers create AI agent Noise2Noise that can denoise images
Adobe to spot fake images using Artificial Intelligence
How Google’s DeepMind is creating images with artificial intelligence

Content Marketing Editor at Packt Hub. I blog about new and upcoming tech trends ranging from Data science, Web development, Programming, Cloud & Networking, IoT, Security and Game development.


Please enter your comment!
Please enter your name here