Trainable speaker diarization

@inproceedings{Aronowitz2007TrainableSD,
  title={Trainable speaker diarization},
  author={Hagai Aronowitz},
  booktitle={INTERSPEECH},
  year={2007}
}
This paper presents a novel framework for speaker diarization. We explicitly model intra-speaker inter-segment variability using a speaker-labeled training corpus and use this modeling to assess the speaker similarity between speech segments. Modeling is done by embedding segments into a segment-space using kernel-PCA, followed by explicit modeling of speaker variability in the segment-space. Our framework leads to a significant improvement in diarization accuracy. Finally, we present a similar… CONTINUE READING