INTERSPEECH 2012
13th Annual Conference of the International Speech Communication Association

Portland, OR, USA
September 9-13, 2012

Phone Adaptive Training for Speaker Diarization

Simon Bozonnet, Ravichander Vipperla, Nicholas Evans

EURECOM, Sophia Antipolis, France

The linguistic content of a speech signal is a source of unwanted variation which can degrade speaker diarization performance. This paper presents our latest work to reduce its impact. The new approach, referred to as Phone Adaptive Training (PAT), is analogous to speaker adaptive training used in automatic speech recognition. We report an oracle experiment which shows that PAT has the potential to deliver a 33% relative improvement in the diarization error rate of our baseline system. Practical experiments show significant improvements across two standard, independent evaluation datasets.

Index Terms: Speaker Diarization, Phone Adaptive Training, Speaker Discrimination

Full Paper

Bibliographic reference.  Bozonnet, Simon / Vipperla, Ravichander / Evans, Nicholas (2012): "Phone adaptive training for speaker diarization", In INTERSPEECH-2012, 494-497.