ML Research Wiki / Benchmarks / Pose Estimation / OCHuman

OCHuman

Pose Estimation Benchmark

Performance Over Time

📊 Showing 18 results | 📏 Metric: Test AP

Top Performing Models

Rank	Model	Paper	Test AP	Date	Code
1	ViTPose (ViTAE-G, GT bounding boxes) 📚	ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation	93.30	2022-04-26	📦 huggingface/transformers 📦 vitae-transformer/vitpose 📦 vitae-transformer/qformer
2	UniHCP (direct eval)	UniHCP: A Unified Model for Human-Centric Perceptions	87.40	2023-03-06	📦 opengvlab/unihcp
3	PoseBH-H 📚	PoseBH: Prototypical Multi-Dataset Training Beyond Human Pose Estimation	87.00	2025-05-23	📦 uyoung-jeong/PoseBH
4	RTMPose(RTMPose-l, GT bounding boxes) 📚	RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose	80.30	2023-03-13	📦 open-mmlab/mmpose
5	TransPose-H	TransPose: Keypoint Localization via Transformer	62.30	2020-12-28	📦 yangsenius/TransPose
6	BBox-Mask-Pose 2x 📚	Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle	48.30	2024-12-02	📦 MiraPurkrabek/BBoxMaskPose
7	BUCTD (CID-W32)	Rethinking pose estimation in crowds: overcoming the detection information-bottleneck and ambiguity	47.20	2023-06-13	📦 amathislab/BUCTD
8	HQNet (ViT-L)	You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception	45.60	2023-12-09	📦 lishuhuai527/coco-unihuman
9	MaskPose-b 📚	Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle	45.00	2024-12-02	📦 MiraPurkrabek/BBoxMaskPose
10	MIPNet (HRNet-W48)	Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation	42.50	2021-01-27	📦 rawalkhirodkar/MIPNet

All Papers (18)

ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation

2022

ViTPose (ViTAE-G, GT bounding boxes)

huggingface/transformers vitae-transformer/vitpose

UniHCP: A Unified Model for Human-Centric Perceptions

2023

UniHCP (direct eval)

opengvlab/unihcp

PoseBH: Prototypical Multi-Dataset Training Beyond Human Pose Estimation

2025

PoseBH-H

uyoung-jeong/PoseBH

RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose

2023

RTMPose(RTMPose-l, GT bounding boxes)

open-mmlab/mmpose

TransPose: Keypoint Localization via Transformer

2020

TransPose-H

yangsenius/TransPose

Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle

2024

BBox-Mask-Pose 2x

MiraPurkrabek/BBoxMaskPose

Rethinking pose estimation in crowds: overcoming the detection information-bottleneck and ambiguity

2023

BUCTD (CID-W32)

amathislab/BUCTD

You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception

2023

HQNet (ViT-L)

lishuhuai527/coco-unihuman

Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle

2024

MaskPose-b

MiraPurkrabek/BBoxMaskPose

Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation

2021

MIPNet (HRNet-W48)

rawalkhirodkar/MIPNet

You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception

2023

HQNet (ResNet-50)

lishuhuai527/coco-unihuman

Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation

2021

HRNet-W48

rawalkhirodkar/MIPNet

Differentiable Hierarchical Graph Grouping for Multi-Person Pose Estimation

2020

HGG (AE+)

Simple Baselines for Human Pose Estimation and Tracking

2018

ResNet-152

PaddlePaddle/PaddleDetection open-mmlab/mmpose

Associative Embedding: End-to-End Learning for Joint Detection and Grouping

2016

Associative Embedding+

open-mmlab/mmpose princeton-vl/pose-ae-train

RMPE: Regional Multi-person Pose Estimation

2016

RMPE

MVIG-SJTU/AlphaPose osmr/imgclsmob

Simple Baselines for Human Pose Estimation and Tracking

2018

ResNet-50

PaddlePaddle/PaddleDetection open-mmlab/mmpose

Associative Embedding: End-to-End Learning for Joint Detection and Grouping

2016

Associative Embedding

open-mmlab/mmpose princeton-vl/pose-ae-train

OCHuman

Performance Over Time

Edit Benchmark Results

Edit Result

Top Performing Models

All Papers (18)

ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation

UniHCP: A Unified Model for Human-Centric Perceptions

PoseBH: Prototypical Multi-Dataset Training Beyond Human Pose Estimation

RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose

TransPose: Keypoint Localization via Transformer

Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle

Rethinking pose estimation in crowds: overcoming the detection information-bottleneck and ambiguity

You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception

Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing the Virtuous Circle

Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation

You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception

Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation

Differentiable Hierarchical Graph Grouping for Multi-Person Pose Estimation

Simple Baselines for Human Pose Estimation and Tracking

Associative Embedding: End-to-End Learning for Joint Detection and Grouping

RMPE: Regional Multi-person Pose Estimation

Simple Baselines for Human Pose Estimation and Tracking

Associative Embedding: End-to-End Learning for Joint Detection and Grouping

Model	Paper	Test AP	Date
ViTPose (ViTAE-G, GT bounding boxes)	ViTPose: Simple Vision Transformer Baselines for …	93.30	2022-04-26
UniHCP (direct eval)	UniHCP: A Unified Model for Human-Centric Percept…	87.40	2023-03-06
PoseBH-H	PoseBH: Prototypical Multi-Dataset Training Beyon…	87.00	2025-05-23
RTMPose(RTMPose-l, GT bounding boxes)	RTMPose: Real-Time Multi-Person Pose Estimation b…	80.30	2023-03-13
TransPose-H	TransPose: Keypoint Localization via Transformer	62.30	2020-12-28
BBox-Mask-Pose 2x	Detection, Pose Estimation and Segmentation for M…	48.30	2024-12-02
BUCTD (CID-W32)	Rethinking pose estimation in crowds: overcoming …	47.20	2023-06-13
HQNet (ViT-L)	You Only Learn One Query: Learning Unified Human …	45.60	2023-12-09
MaskPose-b	Detection, Pose Estimation and Segmentation for M…	45.00	2024-12-02
MIPNet (HRNet-W48)	Multi-Instance Pose Networks: Rethinking Top-Down…	42.50	2021-01-27
HQNet (ResNet-50)	You Only Learn One Query: Learning Unified Human …	40.00	2023-12-09
HRNet-W48	Multi-Instance Pose Networks: Rethinking Top-Down…	37.20	2021-01-27
HGG (AE+)	Differentiable Hierarchical Graph Grouping for Mu…	36.00	2020-07-23
ResNet-152	Simple Baselines for Human Pose Estimation and Tr…	33.30	2018-04-17
Associative Embedding+	Associative Embedding: End-to-End Learning for Jo…	32.80	2016-11-16
RMPE	RMPE: Regional Multi-person Pose Estimation	30.70	2016-12-01
ResNet-50	Simple Baselines for Human Pose Estimation and Tr…	29.50	2018-04-17
Associative Embedding	Associative Embedding: End-to-End Learning for Jo…	29.50	2016-11-16