Computer Graphics, Computer Vision and User Interfaces


Rob Fergus   Davi Geiger   Yann LeCun   Kenneth Perlin      Chee K. Yap   Denis Zorin  

Researchers in Computer Graphics work on computational and mathematical techniques for creating and manipulating computer representations of real and virtual objects and making images of such objects. The main directions of computer graphics research at NYU include animation, geometric modeling, physically-based simulation and computational photography.

The area of Computer Vision is concerned with algorithms and theory necessary to extract information from visual data (images, video, range scans, stereo images, 3D MRI and CAT scan data etc). There is a growing overlap between computer vision and vision research, as the data acquired from images and video is increasingly used in computer graphics applications.

Rob Fergus works in computer vision, machine learning and computer graphics. His goal is to build statistical models of images both at the high level of objects and scenes as well as at the low level of pixels and edges. Such models may then be used for a variety of applications including object recognition, image search and computational photography.

Davi Geiger works in computer vision and on related problems in cognitive science. His current research includes understanding stereo vision, human tracking, shape analysis, and memory structure of vision objects. Much of recent is based on applying Bayesian belief propagation networks (graphical models).

Yann LeCun's research interests include the fundamental and practical aspects of machine learning. The main goal of his research is to devise methods through which computers can extract knowledge and automatically acquire "skills" from massive datasets or from experience. Application of learning to perception is a central theme of his work: how can we teach machines to detect and recognize everyday objects in images, how to teach robots to navigate and avoid obstacles solely from visual input. LeCun's group works on a number of fundamental techniques (energy-based models, "deep learning", relational graphical models, and others) which are applied solve problems in computer vision, robotics, image and signal processing, bioinformatics, medical informatics, and economics. LeCun's group also works with the Center for Neural Science on computational models of biological learning.

Ken Perlin's research spans computer graphics rendering, modeling and animation, user interface software and hardware, surface reflectance measurement devices, and novel display devices. His recent research projects include a collaboration with the New York Hall of Science on creating large scale collaborative educational interfaces for kids, uses of novel interfaces (camera-tracked LED pointers and the Nintendo WiiMote game interface) for collaboration and instruction, and a kaleidoscope-based reflectance measuring device for rapid acquisition of surface reflection functions.

Chee Yap works in the areas of computational geometry, algebraic computation and visualization. His current research focuses on two themes: exact arithmetic computation and large-scale visualization. Numerical nonrobustness has been called "computer scientists' dirty secret" --everyone knows that our numerical software can break, but we rarely talk about it. The most successful approach to solving the robustness problem in geometry is based on exact geometric computation. Chee Yap and his group are developing techniques for this type of computation based on algebraic zero bounds. Exact geometric computation provides considerable insight into the theory of real computation. The other direction of Chee Yap's work is large-scale visualization, dynamic visualization of geospatial data in particular.

Denis Zorin's interests span two areas: geometric modeling, with applications in computer graphics and computer-aided geometric design, and scientific computing and numerical algorithms. In geometric modeling, he works on efficient and accurate discretizations of various types of surface optimization problems, high-order representations for surfaces, including manifold-based and subdivision surfaces and applications to interactive surface and mesh manipulation. In numeral algorithms, he is interested in fast algorithms for solving boundary integral equations, arising from linear PDEs and their applications. Current efforts include a a general fast multipole-based solver for 3D nonhomogeneous PDEs and arbitrary geometries, and on fluid-deformable surface interactions for Stokes fluid.

Participating Faculty from other Departments

Eero Simoncelli


Related Web Pages

Vision Learning Graphics   Media Research Laboratory   Movement Group   Active Visualization   Computational and Biological Learning Lab   Center for Neural Science  

top | contact webmaster