
The proposed D2ST-Adapter is a novel adapter tuning framework that can efficiently and effectively adapt large pre-trained vision models to few-shot action recognition tasks by encoding spatial and temporal features in a disentangled manner using anisotropic deformable spatio-temporal attention.


coremsg

disentangled-and-deformable-spatio-temporal-adapter-for-efficient-few-shot-action-recognition


Disentangled-and-Deformable Spatio-Temporal Adapter for Efficient Few-shot Action Recognition


title_rewrite


The proposed Multi-Velocity Progressive-Alignment (MVP-shot) framework learns and aligns multi-velocity action features to enable more robust and accurate few-shot action recognition, outperforming existing state-of-the-art methods.


multi-velocity-progressive-alignment-framework-for-improved-few-shot-action-recognition


Multi-Velocity Progressive-Alignment Framework for Improved Few-Shot Action Recognition
