Combining larger LLM decoders, preceding-text context, and iterative self-correction reduces character error rate on CNVSRC.Single Chinese visual speech recognition from 49.32% to 38.18%.
Typically, a VSR system takes a silent video containing the speaker’s lip movements as input and outputs the corresponding text
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Leveraging Large Language Models in Visual Speech Recognition: Model Scaling, Context-Aware Decoding, and Iterative Polishing
Combining larger LLM decoders, preceding-text context, and iterative self-correction reduces character error rate on CNVSRC.Single Chinese visual speech recognition from 49.32% to 38.18%.