Diarization partitions an audio stream into segments based on the voices of the speakers. Real-time diarization systems that include an enrollment step should limit enrollment training samples to reduce user interaction time. Although training on a small number of samples yields poor performance, we show that the accuracy can be improved dramatically using a chronological self-training approach. We studied the tradeoff between training time and classification performance and found that 1 second is sufficient to reach over 95% accuracy. We evaluated on 700 audio conversation files of about 10 minutes each from 6 different languages and demonstrated average diarization error rates as low as 10%.
Cite as: Padfield, D., Liebling, D.J. (2021) Chronological Self-Training for Real-Time Speaker Diarization. Proc. Interspeech 2021, 4613-4617, doi: 10.21437/Interspeech.2021-822
@inproceedings{padfield21_interspeech, author={Dirk Padfield and Daniel J. Liebling}, title={{Chronological Self-Training for Real-Time Speaker Diarization}}, year=2021, booktitle={Proc. Interspeech 2021}, pages={4613--4617}, doi={10.21437/Interspeech.2021-822} }