Reverberation-Robust Online Multi-Speaker Tracking by Using a Microphone Array and CASA Processing

Inhalt:
Online tracking of speakers is an important task for applications in smart environments such as camera control, meeting annotation and speech separation. Challenges for an audio-only system are small-room reverberation, noise, the unknown number of speakers, and gaps occurring in natural speech. Combining models from neurobiology and cognitive psychology with many-channel signal processing and pattern recognition techniques, a hybrid method was developed. By employing online CASA processing to signals from a microphone array, the real-time capable method is able to track an arbitrary number of concurrent moving speakers in highly reverberant environments. Index Terms — microphone array, auditory scene analysis, glimpsing model, speaker tracking