Automatic Forensic Voice Comparison Using Recording Adapted Background Models

Becker T, Jessen M, Alsbach S, Broß F, Meier T. Automatic Forensic Voice Comparison Using Recording Adapted Background Models. In: AES Conference: 39th International Conference: Audio Forensics: Practices and Challenges. Audio Engineering Society; 2010. Paper 6-2. Available from: https://aes.org/publications/elibrary-page/?id=15497

@inproceedings{Becker2010_15497,
  author = {Becker, Timo and Jessen, Michael and Alsbach, Sebastian and Broß, Franz and Meier, Torsten},
  title = {{Automatic Forensic Voice Comparison Using Recording Adapted Background Models}},
  booktitle = {AES Conference: 39th International Conference: Audio Forensics: Practices and Challenges},
  note = {Paper 6-2},
  year = {2010},
  month = jun,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=15497}
}

TY  - CPAPER
TI  - Automatic Forensic Voice Comparison Using Recording Adapted Background Models
AU  - Becker, Timo
AU  - Jessen, Michael
AU  - Alsbach, Sebastian
AU  - Broß, Franz
AU  - Meier, Torsten
T2  - AES Conference: 39th International Conference: Audio Forensics: Practices and Challenges
M1  - Paper 6-2
PY  - 2010
DA  - 2010/06/06
UR  - https://aes.org/publications/elibrary-page/?id=15497
PB  - Audio Engineering Society
LA  - en
AB  - We present the automatic forensic voice comparison system SPES which is developed in cooperation with the Federal Criminal Police Office of Germany (Bundeskriminalamt - BKA), the University of Applied Sciences Koblenz and the Department of Phonetics, University of Trier, Germany. The system is based on the classical GMM-UBM framework based on MAP adaption as described by Reynolds et al. This GMM-UBM framework is extended by creating specific background models for each suspect recording based on the similarity of speakers. The less general and more specific background models lead to better discrimination ability and more reliable results. We show how the usage of recording adapted background models outperforms the standard GMM-UBM approach in different recording conditions.
ER  -