Cornell is a leader in computer graphics, an interdisciplinary area that draws on many specialties including algorithms, physics, computation, psychology, computer vision, and architecture. The Cornell graphics tradition has roots going back to the earliest days of the field, when the Program of Computer Graphics (PCG) was established in 1974 and went on to make breakthrough contributions in areas including light reflection models, physics-based accurate rendering, and visual perception for graphics. Today graphics research at Cornell flows across boundaries to cover a broad area of graphics and related topics, with research in graphics and vision in the Computer Science department, research in graphics and architecture in PCG, and research in human-computer interfaces in the Information Science program, all densely interconnected.

Current research in graphics covers a broad range of topics across the field. Examples include global illumination, scattering models, volume scattering, interactive rendering, rigid and flexible body simulation, cloth simulation, acoustics for graphics, haptics, multiview geometry, computational photography, human visual perception, and appearance capture. Our research addresses applications ranging from film effects, animation, and games to architecture, surgery simulation, cosmetics, and photo browsing.

Faculty: Research Descriptions

Kavita Bala works on computer graphics algorithms for rendering and modeling complex virtual worlds. A fundamental challenge is efficiently capturing the visual complexity and richness of real scenes. By understanding and exploiting the limits of the human visual system, new rendering and modeling algorithms become possible that scale to real-world complexity. Bala's research interests include scalable graphics for interactive rendering; perceptually-based rendering; image-based modeling and texturing; cinematic relighting; and feature-based rendering and texturing. Applications range from cultural heritage and preservation, engineering design, games and movies, virtual-reality training, architectural planning, and e-commerce.

Donald Greenberg, the founder of the Program of Computer Graphics, has been researching and teaching in the field of computer graphics from 1966. During the last 15 years, he has been primarily concerned with research advancing the state-of-the-art in computer graphics and with utilizing these techniques as they may be applied to a variety of disciplines. His specialities include hidden surface algorithms, geometric modeling, color science, and realistic image generation. Donald Greenberg is the Jacob Gould Schurman Professor of Computer Graphics and the Director of the Program of Computer Graphics.

Doug James works on physics-based algorithms for computer graphics, physically based animation, haptics and sound rendering, and scientific computing. Some research themes are (1) efficient deformation processing; (2) interactive and multi-sensory (graphics, haptics, sound) physical simulation; and (3) motion control and design. Designing algorithms to exploit the structure and information content of physical phenomena can permit faster and better simulations. An important research theme has been the design of amortized algorithms that leverage preprocessing for faster physical simulations. Some current research addresses how to accelerate processing of discrete deformable systems: fast subspace integration of solid dynamics, output-sensitive collision detection techniques, fast contact resolution, force-feedback haptic rendering, and physically based sound rendering.

Steve Marschner works on material models and model capture, often using techniques that draw from computer vision. Material modeling is the fundamental problem of understanding and simulating the interaction of light with materials. Existing models describe simple materials well, but we need new insights before we can accommodate the complex mixtures of materials that occur commonly in reality. Model capture focuses on developing methods for building richly detailed models using measurements from real objects. This work is part of a major trend that is increasingly blurring the distinction between vision and graphics: more and more graphics applications have extended the limits of traditional modeling by using a mixture of vision and graphics techniques to import complexity from the real world.

Noah Snavely is primarily interested in analyzing large image collections to automatically recover the geometry and appearance of real-world scenes, and in using this derived structure to create better visualizations of photo collections and 3D scenes. Noah is particularly interested in leveraging the vast, rich collections of imagery available on the Internet to recreate the world in 3D. This research encompasses problems in both computer vision and computer graphics, including structure from motion, multi-view stereo, graph algorithms for analyzing large image collections, image-based rendering, and 3D navigation interfaces. Noah is also interested in creating systems and techniques that make it simple to create 3D models using a hand-held camera. He will be joining the department in January of 2009.

Program of Computer Graphics

Researchers

Kavita Bala
Don Greenberg
Doug James
Steve Marschner
Noah Snavely
Ken Torrance
Bruce Walter

Related Areas

Scientific and Parallel Computing
Artificial Intelligence

Related researchers

Paul Chew
Jim Ferwerda
Franois Guimbretire
Dan Huttenlocher
Ramin Zabih