Phoneme-level decomposition of speech improves person-of-interest deepfake detection by making decisions robust to noise and compression while localizing artifacts to specific phonemes.
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
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Phoneme-Level Analysis for Person-of-Interest Speech Deepfake Detection
Phoneme-level decomposition of speech improves person-of-interest deepfake detection by making decisions robust to noise and compression while localizing artifacts to specific phonemes.