TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/DGNN

DGNN

Reported on 32 benchmarks across 8 tasks

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision12 results

  • VideoonKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • VideoonUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • VideoonNTU RGB+D
    Accuracy (CS)
    89.9
    best: 94.3 (Hulk(Finetune, ViT-L))
  • VideoonNTU RGB+D
    Accuracy (CV)
    96.1
    best: 98.3 (ST-GCN [PYSKL, 2D Skeleton])
  • Temporal Action LocalizationonKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • Temporal Action LocalizationonUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • Temporal Action LocalizationonNTU RGB+D
    Accuracy (CS)
    89.9
    best: 94.3 (Hulk(Finetune, ViT-L))
  • Temporal Action LocalizationonNTU RGB+D
    Accuracy (CV)
    96.1
    best: 98.3 (ST-GCN [PYSKL, 2D Skeleton])
  • Action LocalizationonKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • Action LocalizationonUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • Action LocalizationonNTU RGB+D
    Accuracy (CS)
    89.9
    best: 94.3 (Hulk(Finetune, ViT-L))
  • Action LocalizationonNTU RGB+D
    Accuracy (CV)
    96.1
    best: 98.3 (ST-GCN [PYSKL, 2D Skeleton])

Time Series8 results

  • Action DetectiononKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • Action DetectiononUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • Action DetectiononNTU RGB+D
    Accuracy (CS)
    89.9
    best: 94.3 (Hulk(Finetune, ViT-L))
  • Action DetectiononNTU RGB+D
    Accuracy (CV)
    96.1
    best: 98.3 (ST-GCN [PYSKL, 2D Skeleton])
  • Action RecognitiononKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • Action RecognitiononUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • Action RecognitiononNTU RGB+D
    Accuracy (CS)
    89.9
    best: 97.4 (DSCNet (RGB + Pose))
  • Action RecognitiononNTU RGB+D
    Accuracy (CV)
    96.1
    best: 99.6 (PoseC3D (RGB + Pose))

Methodology4 results

  • Zero-Shot LearningonKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • Zero-Shot LearningonUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • Zero-Shot LearningonNTU RGB+D
    Accuracy (CS)
    89.9
    best: 94.3 (Hulk(Finetune, ViT-L))
  • Zero-Shot LearningonNTU RGB+D
    Accuracy (CV)
    96.1
    best: 98.3 (ST-GCN [PYSKL, 2D Skeleton])

Robots4 results

  • Activity RecognitiononKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • Activity RecognitiononUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • Activity RecognitiononNTU RGB+D
    Accuracy (CS)
    89.9
    best: 97.4 (DSCNet (RGB + Pose))
  • Activity RecognitiononNTU RGB+D
    Accuracy (CV)
    96.1
    best: 99.6 (PoseC3D (RGB + Pose))

Natural Language Processing4 results

  • 3D Action RecognitiononKinetics-Skeleton dataset
    Accuracy
    36.9
    best: 52.3 (Structured Keypoint Pooling (PPNv2 skeletons+objects))
  • 3D Action RecognitiononUAV-Human
    CSv1(%)
    29.9
    best: 47.96 (HDBN)
  • 3D Action RecognitiononNTU RGB+D
    Accuracy (CS)
    89.9
    best: 94.3 (Hulk(Finetune, ViT-L))
  • 3D Action RecognitiononNTU RGB+D
    Accuracy (CV)
    96.1
    best: 98.3 (ST-GCN [PYSKL, 2D Skeleton])