BHaRNet: Reliability-Aware Body-Hand Modality Expertized Networks for Fine-grained Skeleton Action Recognition

Open paper
Benchmark Model Rank Results
skeleton-based-action-recognition-on-ntu-rgbd-120BHaRNet-B–Accuracy (Cross-Subject): 94.5Accuracy (Cross-Setup): 95.5
skeleton-based-action-recognition-on-ntu-rgbd-120BHaRNet-E–Accuracy (Cross-Subject): 94.7Accuracy (Cross-Setup): 95.7
skeleton-based-action-recognition-on-ntu-rgbd-120BHaRNet-P–Accuracy (Cross-Subject): 94.8Accuracy (Cross-Setup): 95.8
skeleton-based-action-recognition-on-ntu-rgbd-60BHaRNet-B–Accuracy (CS): 96.6Accuracy (CV): 99.1
skeleton-based-action-recognition-on-ntu-rgbd-60BHaRNet-E–Accuracy (CS): 96.7Accuracy (CV): 99.2
skeleton-based-action-recognition-on-ntu-rgbd-60BHaRNet-P–Accuracy (CS): 96.8Accuracy (CV): 99.2

Rank counts only results with a code link.