ISCA Archive Interspeech 2008
ISCA Archive Interspeech 2008

Modulation spectrogram features for improved speaker diarization

Oriol Vinyals, Gerald Friedland

We propose the use of modulation spectrogram features in speaker diarization. These features carry longer term characteristics of the acoustic signals than the widely used MFCCs, thus providing potential improvement by using both features in combination. Using the state-of-the-art ICSI speaker diarization system, an improvement of 20.77% relative DER is obtained on the NIST Rich Transcription 2007 task with respect to the MFCC only system.


doi: 10.21437/Interspeech.2008-199

Cite as: Vinyals, O., Friedland, G. (2008) Modulation spectrogram features for improved speaker diarization. Proc. Interspeech 2008, 630-633, doi: 10.21437/Interspeech.2008-199

@inproceedings{vinyals08_interspeech,
  author={Oriol Vinyals and Gerald Friedland},
  title={{Modulation spectrogram features for improved speaker diarization}},
  year=2008,
  booktitle={Proc. Interspeech 2008},
  pages={630--633},
  doi={10.21437/Interspeech.2008-199}
}