Joint Audio-Video Object Tracking

Spors S, Rabenstein R, Strobel N (2001)


Publication Status: Published

Publication Type: Conference contribution

Publication year: 2001

Journal

Publisher: Institute of Electrical and Electronic Engineers

Book Volume: 18

Pages Range: 22-31

Event location: Thessaloniki GR

Journal Issue: 1

URI: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=0035075362&origin=inward

Abstract

A method for joint localization of a single object using an audio sensor and a video sensor were outlined. These localizers are best described using different coordinate systems as the audio position estimates are nonlinearly related to the state. Whereas video sensor system operates in Cartesian coordinates. The local Kalman filters consists of three main components: local Kalman filter at the microphone array; a local Kalman filter at the video camera; and a global Kalman filter at the fusion center. The loss of optimality is the price to pay for replacing the linear Kalman filter with an extended Kalman filter. Quantitative experiments have shown that such a strategy can improve object localization and tracking.

Authors with CRIS profile

How to cite

APA:

Spors, S., Rabenstein, R., & Strobel, N. (2001). Joint Audio-Video Object Tracking. In Proceedings of the IEEE International Conference on Image Processing (ICIP) (pp. 22-31). Thessaloniki, GR: Institute of Electrical and Electronic Engineers.

MLA:

Spors, Sascha, Rudolf Rabenstein, and Norbert Strobel. "Joint Audio-Video Object Tracking." Proceedings of the IEEE International Conference on Image Processing (ICIP), Thessaloniki Institute of Electrical and Electronic Engineers, 2001. 22-31.

BibTeX: Download