Seminar Homepage
This Week's Seminar
Last Week's Seminar
Seminar Archive
Fall 1998 Schedule

Seminar Archive

Warning
This page is provided for historical and archival purposes only. While the seminar dates are correct, we offer no guarantee of informational accuracy or link validity. Contact information for the speakers, hosts and seminar committee are certainly out of date.

Toward Image-Based Modeling of Complex Environments

Steve Seitz, Assistant Professor
Robotics Institute and Computer Science
Carnegie Mellon University

Place and Time
Adamson Wing, Baker Hall
Refreshments 3:45 pm
Talk 4:00 pm

Abstract
The problem of acquiring graphical models of complex environments is a research topic with far-ranging significance in computer vision, graphics, and virtual reality. The primary goal is to model a real environment, such as a landscape or a building exterior and interior, in a manner that enables storing, transmitting, and rendering that environment from any viewpoint, with photographic quality and detail. Recent interest in this problem has been driven by the proliferation of high-quality digital imagery from which such models may be constructed, the emergence of powerful 3D graphics capabilities on desktop PC's, and numerous potential applications in such areas as virtual environments, video teleconferencing, and web-based commerce.

In this talk I will describe open research problems and present new results aimed at reconstructing complex objects and scenes from photographs. Chief among these problems are 1) visibility: determining which portions of the (unknown) environment are visible in each input image, and 2) correspondence: integrating measurements from numerous input images. The talk will introduce a powerful class of algorithms for solving visibility and correspondence problems under very general conditions. Based on these results, we have developed new voxel- based reconstruction algorithms for computing photorealistic 3D models of real environments. Results will be presented for a range of different scenes and image sequences.

Speaker Biography
Steve Seitz is an Assistant Professor in the Robotics Institute at Carnegie Mellon University. Before joining the institute in August 1998, he spent a year visiting the Vision Technology Group at Microsoft Research, where he conducted research in image-based rendering and visual motion analysis. His current research focuses on the problem of acquiring and manipulating visual representations of real environments using semi- and fully-automated techniques. He received his BA in computer science and mathematics at the University of California, Berkeley in 1991, his Ph.D. in computer science at the University of Wisconsin--Madison in 1997, and worked as a summer intern in the Advanced Technology Group at Apple Computer in 1993.

Speaker Appointments
For appointments, please contact the speaker, Steve Seitz, seitz@ri.cmu.edu.