Interact before Align: Leveraging Cross-Modal Knowledge for Domain Adaptive Action Recognition

Publication
Proceedings of the 2022 Conference on Computer Vision and Pattern Recognition (CVPR 2022)