This paper studies signal enhancement of microphone array sound pickup and neural network (NN) adaptation for reliable speaker identification. Additionally, the report describes a speaker identification system for teleconferencing applications. It is found that beamforming/matched-filter microphone arrays are capable of combating deleterious properties of the acoustic environment, such as multipath distortion (reverberation) and ambient noise. It is also found that neural networks can learn, and thereby compensate for, environmental interference and can adapt testing conditions to training conditions, so that high accuracies can be achieved in speaker identification.
Cite as: Lin, Q., Jan, E., Che, C., Flanagan, J.L. (1994) Speaker identification in teleconferencing environments using microphone arrays and neural networks. Proc. ESCA Workshop on Automatic Speaker Recognition, Identification and Verification, 235-238
@inproceedings{lin94_asriv,
author={Q. Lin and E. Jan and C. Che and James L. Flanagan},
title={{Speaker identification in teleconferencing environments using microphone arrays and neural networks}},
year=1994,
booktitle={Proc. ESCA Workshop on Automatic Speaker Recognition, Identification and Verification},
pages={235--238}
}