Cross-Modal Learning with 3D Deformable Attention for Action Recognition

Benchmark Model Rank Results
action-recognition-in-videos-on-ntu-rgbd-1203DA (RGB + Pose)Accuracy (Cross-Setup): 91.4Accuracy (Cross-Subject): 90.5
action-recognition-in-videos-on-ntu-rgbd-603DA (RGB + Pose)Accuracy (CS): 94.3Accuracy (CV): 97.9