Chronological Self-Training for Real-Time Speaker Diarization

Autor:	Padfield, Dirk, Liebling, Daniel J.
Rok vydání:	2022
Předmět:	Computer Science - Sound Computer Science - Computation and Language Computer Science - Machine Learning Electrical Engineering and Systems Science - Audio and Speech Processing
Zdroj:	Proc. Interspeech (2021) 4613-4617
Druh dokumentu:	Working Paper
DOI:	10.21437/Interspeech.2021-822
Popis:	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%. Comment: 5 pages, 5 figures, ICASSP 2021
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2208.03393 Zobrazit plný text záznamu View this record from Arxiv