asvspoof / D3M

D3M - Dynamic Data Discrepancy Mitigation for Anti-spoofing - Implementation of work Dynamically Mitigating Data Discrepancy with Balanced Focal Loss for Replay Attack Detection
https://arxiv.org/abs/2006.14563v3
MIT License
28 stars 4 forks source link

Is it working good in out of domain data? #10

Closed mesut92 closed 2 years ago

mesut92 commented 3 years ago

Hi; Did you try this real-world data? Or do you have any suggestions?

Thanks, Mesut

DeeDive commented 3 years ago

Hi @mesut92, as we note in the paper, it does not work well in the real-world dataset, which might be a crucial direction for future research. And from my perspective, you can also try cross-dataset validation.