Subband based voice spoofing detection

In this work, Bhusan Chettri explores the use of individual sub-bands (1 kHz bands) towards voice spoofing detection tasks and further compares how performance differs from models trained on the whole frequency bands.

Transcript

2. Automatic speaker verification (ASV) Is the speaker who he/she claims to be?
3. Spoofing and countermeasures
4. Spoofing and countermeasures
5. Fullband countermeasure
6. Fullband countermeasures: baselines
7. Proposed methodology
8. Proposed methodology
9. Fullband baseline results [1] B. Chettri et. al. Ensemble Models for Spoofing Detection in Automatic Speaker Verification. Proc. Interspeech 2019. [2] H. Delgado et. al. ASVspoof 2017 Version 2.0: meta-data analysis and baseline enhancements. Speaker Odyssey 2018. [3] M. Todisco et. al. ASVspoof 2019: Future Horizons in Spoofed and Fake Audio Detection. Proc. Interspeech 2019.
10. 4 kHz subband CNN results
11. 4 kHz subband CNN results
12. 2 kHz subband CNN results
13. 2 kHz subband CNN results
14. 1 kHz subband CNN results
16. Summary and conclusions