Bridging the Skeleton-Text Modality Gap: Diffusion-Powered Modality Alignment for Zero-Shot Skeleton-Based Action Recognition.

Jeonghyeok Do, Munchurl Kim

Browse the full ICCV paper archive.