Research Projects: Automatic Creation of Three-Dimensional Avatars
|Line 18:||Line 18:|
== External Links ==
== External Links ==
Revision as of 02:37, 12 February 2006
Project Name: Automatic Creation of Three-Dimensional Avatars
Team Members: Maria-Cruz Villa-Uriol, Falko Kuester and Nader Bagherzadeh
Project Sponsor: National Science Foundation (NSF) and Balsells Fellowship
Traditionally, the acquisition of three-dimensional avatars from real humans relied on the utilization of body suits, optical or magnetic markers, or complex and expensive three-dimensional scanners. In general, entertainment companies, especially those dedicated to computer games and computer-animated movies, have invested into and leveraged from these techniques. Some of the main drawbacks that these capture systems introduce are equipment cost, portability and the required degree of human intervention to ensure compelling visual results. Alternative techniques have been investigated by the computer vision community, aimed at reconstructing human body models from video sequences or sets of static images. However, the primary focus of these techniques generally was on motion analysis and tracking whereas reconstruction of realistic avatars was only a secondary objective. This project consists of developing an avatar construction pipeline designed to use multiple standard video cameras for the creation of realistic three-dimensional avatars that can be included into interactive virtual environments. Multiple images of a human subject are used to automatically reshape a synthetic three-dimensional articulated reference model into a high-quality avatar. The pipeline under development combines software and hardware-accelerated stages into one seamless system. Primary stages in this pipeline include pose estimation, skeleton fitting, body part segmentation, geometry construction and texturing. Silhouette-based modification techniques in combination with a reference model remove traditional constraints in the initial pose of the captured subject. Results can be obtained in near-real time with very limited user intervention.
Related Resarch Projects
Tools Used In This Project
This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by the author's copyright. This work may not be reposted without the explicit permission of the copyright holder.