AI, Computer Graphics, and Design at Siggraph

The fusion of artificial intelligence (AI) with computer graphics is revolutionizing the field, delivering unprecedented realism and efficiency. AI advancements have significantly improved graphic rendering and animation capabilities. Generative AI and neural rendering are now transforming 2D images into immersive 3D landscapes, achieving remarkable visual fidelity in real-time simulations.

At the forefront of these innovations is CLAY, a cutting-edge model that translates human imagination into digital forms with astonishing accuracy. The model, featured in the research paper “CLAY: A Controllable Large-scale Generative Model for Creating High-quality 3D Assets,” showcases its ability to create detailed 3D assets, as demonstrated by a team from ShanghaiTech and Deemos, alongside Huazhong University of Science and Technology.

Another breakthrough is the DreamMat model, a geometry- and light-aware diffusion model that generates high-quality PBR (physically based rendering) materials. Developed by researchers from Zhejiang University, the University of Hong Kong, and Tencent Games, DreamMat uses text prompts to create realistic object appearances, significantly lowering the barrier for producing high-quality visual assets in movies, games, and AR/VR applications.

Real-time character control has also taken a leap forward with the Conditional Autoregressive Motion Diffusion Model (CAMDM). This model, developed by a team from the Hong Kong University of Science and Technology, the University of Hong Kong, and Tencent AI Lab, generates diverse and realistic character animations in real-time, enhancing the interactive experience in gaming and virtual environments.

These groundbreaking advancements will be showcased at SIGGRAPH 2024, held from July 28 to August 1 in Denver. This year’s conference will highlight significant progress in the integration of AI with computer graphics, heralding a future where digital and real worlds seamlessly converge. Attendees can expect to see detailed presentations of these innovations as part of the Technical Papers program, offering a glimpse into the future of content creation and interactive design.

“Generative AI is set to democratize content creation and transform various industries,” says Qixuan (QX) Zhang, project lead for CLAY. “We foresee a future where advanced 3D content creation and interaction become accessible to everyone, enhancing productivity and creativity across professional domains.”

For more information on these exciting developments and to explore the full lineup of research, visit the SIGGRAPH 2024 Technical Papers program.