Systems Seminar - ECE
New Approaches to Automatic Speaker Recognition and Forensic Considerations
The area of automatic speaker recognition has been dominated by systems using only short-term, low-level acoustic information, such as cepstral features. While these systems have produced reasonably low error rates, they ignore other levels of information beyond low-level acoustics that convey speaker information. This presentation will cover recent research on methods that exploit high-level information. For example, idiosyncratic word usage and pronunciation are high-level features that can be exploited. Furthermore, these high-level features can be combined with conventional features for increased accuracy. New compensation methods will also be covered that increase robustness due to degrading factors common to the forensic domain. The presentation concludes with appropriate uses of this technology, especially cautions regarding forensic-style applications, and looking at its future directions.
Joseph P. Campbell received the B.S. degree in electrical engineering from Rensselaer Polytechnic Institute in 1979, the M.S. degree in electrical engineering from The Johns Hopkins University in 1986, and the Ph.D. degree in electrical engineering from Oklahoma State University in 1992. Joe is currently the Assistant Group Leader in the Human Language Technology Group at MIT Lincoln Laboratory, where he conducts speech-processing research and specializes in speaker recognition, evaluation and corpus design, biometrics, and forensics. Before joining Lincoln Laboratory as a senior member of the technical staff, he served 22 years at the National Security Agency (NSA). He is currently a Chair of the International Speech Communication Association's Speaker and Language Characterization Special Interest Group (ISCA SpLC SIG) and Vice President of Technical Activities of the IEEE Biometrics Council. Dr. Campbell is a member of the National Academy of Sciences' Whither Biometrics? Committee, the IEEE Jack S. Kilby Signal Processing Medal Committee, and the IEEE Information Forensics Security Technical Committee. He is a member of Sigma Xi, the International Speech Communication Association, the Boston Audio Society, and the Acoustical Society of America. Dr. Campbell was named a Fellow of the IEEE "for leadership in biometrics, speech systems, and government applications" in 2005.