SiLVERScore, built on the CiCo video-text contrastive model, discriminates correct vs. random sign-video/text pairs with 0.99 ROC AUC and is robust to word reordering and prosody intensity.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SiLVERScore: Semantically-Aware Embeddings for Sign Language Generation Evaluation
SiLVERScore, built on the CiCo video-text contrastive model, discriminates correct vs. random sign-video/text pairs with 0.99 ROC AUC and is robust to word reordering and prosody intensity.