Learning the Time-Delay Manifold for Robust Speaker Localization

Evan Ettinger, Shankar Shivappa, Deborah Goshorn and Yoav Freund
December 4, 2007

We present an algorithm for high dimensional density estimation which is efficient (both computationally and statistically) when the distribution is concentrated close to a low dimensional smooth manifold. The algorithm uses several random projections to generate a hierarchical mixture of Gaussians which rapidly converges to the underlying manifold. We use this algorithm to perform robust estimation of the time delays in an ad-hoc microphone network. We utilize the model to calculate accurate time-delay vectors for two speakers that are talking at the same time.

How to view this document

The authors of these documents have submitted their reports to this technical report series for the purpose of non-commercial dissemination of scientific work. The reports are copyrighted by the authors, and their existence in electronic format does not imply that the authors have relinquished any rights. You may copy a report for scholarly, non-commercial purposes, such as research or instruction, provided that you agree to respect the author's copyright. For information concerning the use of this document for other than research or instructional purposes, contact the authors. Other information concerning this technical report series can be obtained from the Computer Science and Engineering Department at the University of California at San Diego, techreports@cs.ucsd.edu.

[ Search ]

This server operates at UCSD Computer Science and Engineering.
Send email to webmaster@cs.ucsd.edu