Speech-Forensics combines multi-span partial forgeries with per-span algorithm labels, and the TEST network detects authenticity, localizes fake segments, and identifies synthesis algorithms simultaneously.
wav2vec 2.0: A framework for self-supervised learning of speech representations
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Speech-Forensics: Towards Comprehensive Synthetic Speech Dataset Establishment and Analysis
Speech-Forensics combines multi-span partial forgeries with per-span algorithm labels, and the TEST network detects authenticity, localizes fake segments, and identifies synthesis algorithms simultaneously.