Subband fusion of complex spectrogram for fake speech detection
Speech Communication · 2023-11 · 155 · 102988
DOI: 10.1016/j.specom.2023.102988 ↗Kurz gefasst
Fachartikel in der Zeitschrift Speech Communication, 2023-11. Verfasst von Cunhang Fan und 6 weiteren Autor:innen. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 59 Literaturangaben.
Abstract (Originalsprache)
The phase information was shown useful in fake speech detection. However, the most common reason why phase-based features are not widely used is phase wrapping. This makes the original phase hard to model directly. Therefore, it remains a challenge how to utilize the phase information effectively. To address this issue, this paper proposes a novel subband fusion of the complex spectrogram method for fake speech detection. The complex spectrogram is used as the input feature, containing both amplitude and phase spectrogram. In addition, subbands of the complex spectrogram are modeled separately. The idea is motivated by the fact that each frequency band has a different effect on the fake speech detection task. Finally, to make full use of the subbands, the subband results are fused. Experimental results on the ASVspoof 2019 LA dataset show that our proposed system achieves an equal error rate (EER) of 0.68% and a minimum tandem detection cost function (min t-DCF) of 0.0224.
Zitieren
APA
Fan, C., Xue, J., Dong, S., Ding, M., Yi, J., Li, J., & Lv, Z. (2023). Subband fusion of complex spectrogram for fake speech detection. Speech Communication, 155, 102988. https://doi.org/10.1016/j.specom.2023.102988
Vancouver
Fan C, Xue J, Dong S, Ding M, Yi J, Li J, et al. Subband fusion of complex spectrogram for fake speech detection. Speech Communication. 2023;155:102988. doi:10.1016/j.specom.2023.102988
BibTeX
@article{Fan2023,
author = {Fan, Cunhang and Xue, Jun and Dong, Shunbo and Ding, Mingming and Yi, Jiangyan and Li, Jinpeng and Lv, Zhao},
title = {Subband fusion of complex spectrogram for fake speech detection},
journal = {Speech Communication},
year = {2023},
volume = {155},
pages = {102988},
doi = {10.1016/j.specom.2023.102988}
}
EUVIMED