SIGGRAPH (Special Interest Group on Computer Graphics and Interactive Techniques) is the ACM’s premier annual conference on computer graphics and interactive technology, considered the world’s most prestigious venue for publishing advances in rendering, animation, simulation, geometry processing, and human-computer interaction. First held in 1974, SIGGRAPH has been the introduction point for transformative techniques including ray tracing, physically based rendering, GPU shading languages, neural rendering, and real-time global illumination. Its technical papers programme is among the most selective in computer science, with acceptance rates typically below 25 per cent.
Content
- SIGGRAPH was founded in 1969 as a special interest group within the Association for Computing Machinery, with the first conference held in Boulder, Colorado in 1974. It rapidly became the essential gathering of the computer graphics community, drawing researchers from academia, film studios, game studios, and hardware manufacturers. The conference’s dual identity—rigorous technical papers programme alongside a commercial exhibition and industry presentations—reflects the unusually tight coupling between research and practice in computer graphics.
- The list of techniques first published at SIGGRAPH reads as the history of modern visual computing. Jim Blinn’s bump mapping (1978), Turner Whitted’s recursive ray tracing (1980), the Cook-Torrance BRDF model (1982), Pixar’s RenderMan shading language (1988), the Metropolis light transport algorithm (1997), precomputed radiance transfer (2002), and NVIDIA’s screen-space ambient occlusion (2007) were all introduced at SIGGRAPH. This publication record means that virtually every rendered image—whether in a blockbuster film, a AAA video game, or an architectural visualisation—traces its technical lineage through SIGGRAPH papers.
- The past decade has seen SIGGRAPH transformed by the intersection of deep learning with graphics. Neural radiance fields (NeRF), introduced at ECCV 2020 but rapidly dominating SIGGRAPH papers thereafter, represent a paradigm shift from explicit geometric representations to implicit neural scene representations learned from images. 3D Gaussian Splatting, presented at SIGGRAPH 2023, provided a real-time-capable neural rendering alternative that has been rapidly adopted in robotics, AR/VR, and digital twin applications. These developments demonstrate SIGGRAPH’s continuing role as the bridge between ML research and deployable visual computing.
- SIGGRAPH Asia, launched in 2008, extends the conference’s reach to the Asia-Pacific research and production community, typically alternating between Singapore, Tokyo, and other major cities. The two conferences together publish approximately 400-500 technical papers annually, representing a substantial fraction of the world’s highest-impact computer graphics research output. The SIGGRAPH Digital Library, maintained by ACM, provides open access to this archive through the ACM DL, making the full body of computer graphics knowledge accessible to practitioners worldwide.