Abstract
Synthesis artifacts that span scales from small to large are important cues for spoofing detection. However, few spoofing detection models leverage artifacts across different scales together. In this paper, we propose a spoofing detection system built on SincNet and Deep Layer Aggregation (DLA), which leverages speech representations at different levels to distinguish synthetic speech. DLA is totally convolutional with an iterative tree-like structure. The unique topology of DLA makes possible compounding of speech features from convolution layers at different depths, and therefore the local and the global speech representations can be incorporated simultaneously. Moreover, SincNet is employed as the frontend feature extractor to circumvent manual feature extraction and selection. SincNet can learn fine-grained features directly from the input speech waveform, thus making the proposed spoofing detection system end-to-end. The proposed system outperforms the baselines when tested on ASVspoof LA and DF datasets. Notably, our single model surpasses all competing systems in ASVspoof DF competition with an equal error rate (EER) of 13.99%, which demonstrates the importance of multi-scale information aggregation for synthetic speech detection.
Original language | English |
---|---|
Article number | 57 |
Journal | Eurasip Journal on Audio, Speech, and Music Processing |
Volume | 2024 |
Issue number | 1 |
DOIs | |
Publication status | Published - Dec 2024 |
Externally published | Yes |
Bibliographical note
Publisher Copyright:© The Author(s) 2024.
ASJC Scopus Subject Areas
- Acoustics and Ultrasonics
- Electrical and Electronic Engineering
Keywords
- Convolutional neural network
- Deep fake detection
- Deep learning
- Information aggregation
- Voice anti-spoofing
Press/Media
-
Chinese Academy of Sciences Researchers Have Published New Data on Data Aggregation (Multi-scale Information Aggregation for Spoofing Detection)
11/25/24
1 item of Media coverage
Press/Media: Research