HQNet (ResNet-50)

Reported on 5 benchmarks across 5 tasks · 1 paper

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

Computer Vision3 results

Pose EstimationonOCHuman
Test AP· 2023-12-09
40
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception arXiv:2312.05525
Instance SegmentationonOCHuman
AP· 2023-12-09
31.1
best: 55.2 (Pose2Seg (plus ground-truth keypoints))
You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception arXiv:2312.05525
Human Instance SegmentationonOCHuman
AP· 2023-12-09
31.1
best: 55.2 (Pose2Seg (plus ground-truth keypoints))
You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception arXiv:2312.05525

Methodology1 result

3DonOCHuman
Test AP· 2023-12-09
40
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception arXiv:2312.05525

Audio1 result

1 Image, 2*2 StitchionOCHuman
Test AP· 2023-12-09
40
best: 93.3 (ViTPose (ViTAE-G, GT bounding boxes))
You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception arXiv:2312.05525