Articulation-manner knowledge vectors improve target speech extraction on cinematic audio source separation (CASS) data versus knowledge-free baselines, especially for short buried speech segments.
Fant,Speech Sounds and Features
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Knowledge-Driven Approach to Target Speech Extraction in the Presence of Background Sound Effects for Cinematic Audio Source Separation (CASS)
Articulation-manner knowledge vectors improve target speech extraction on cinematic audio source separation (CASS) data versus knowledge-free baselines, especially for short buried speech segments.