Multiple Input Feature Sets from Real-Time Color and Range Data for Reliable Tracking

Conference · Tue Oct 19 00:00:00 EDT 1999

OSTI ID:791123

Tsap, L V; Goldgof, D B

This paper describes a work in progress on using multiple sets of input features for robust real-time object tracking in image sequences. Traditional approaches to tracking relied mostly on segmentation of the intensity data using motion or appearance data. Recent availability of real-time range data allows us to use it as an additional unrivaled source of information. We propose a combination of intensity- and range-based input features. Range data enables localized search for' specific features which improves tracking reliability and speed. Proposed approach was successfully tested for the face and gesture tracking application.

• Research Organization: Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
• Sponsoring Organization: USDOE Office of Defense Programs (DP) (US)
• DOE Contract Number: W-7405-Eng-48
• OSTI ID: 791123
• Report Number(s): UCRL-JC-136053; TRN: US200513%%185
• Resource Relation: Conference: International Workshop on Digital and Computational Video, Tampa, FL (US), 12/10/1999; Other Information: PBD: 19 Oct 1999
• Country of Publication: United States
• Language: English

Related Subjects

99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE
AVAILABILITY
COLOR
RELIABILITY
VELOCITY