DynaPURLS adapts textual action descriptions at inference time using the model's own confident predictions, improving zero-shot skeleton action recognition accuracy on NTU60/120 and PKU-MMD over static-matching baselines.
Title resolution pending
1 Pith paper cite this work, alongside 34 external citations. Polarity classification is still indexing.
1
Pith paper citing it
34
external citations · OpenAlex
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DynaPURLS: Dynamic Refinement of Part-Aware Representations for Skeleton-Based Zero-Shot Action Recognition
DynaPURLS adapts textual action descriptions at inference time using the model's own confident predictions, improving zero-shot skeleton action recognition accuracy on NTU60/120 and PKU-MMD over static-matching baselines.