PoTion

Reported on 8 benchmarks across 8 tasks

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

Computer Vision3 results

VideoonJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)
Temporal Action LocalizationonJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)
Action LocalizationonJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)

Time Series2 results

Action DetectiononJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)
Action RecognitiononJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)

Methodology1 result

Zero-Shot LearningonJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)

Robots1 result

Activity RecognitiononJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)

Natural Language Processing1 result

3D Action RecognitiononJHMDB (2D poses only)
Average accuracy of 3 splits
67.9
best: 86.1 (HT-ConvNet)