TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
2D Human Pose Estimation	COCO-WholeBody	RTMPose	WB	65.3	# 3
2D Human Pose Estimation	COCO-WholeBody	RTMPose	body	71.4	# 7
2D Human Pose Estimation	COCO-WholeBody	RTMPose	foot	69.2	# 5
2D Human Pose Estimation	COCO-WholeBody	RTMPose	face	88.9	# 1
2D Human Pose Estimation	COCO-WholeBody	RTMPose	hand	59.0	# 4
2D Human Pose Estimation	Human-Art	RTMPose-l	AP	0.378	# 9
2D Human Pose Estimation	Human-Art	RTMPose-l	AP (gt bbox)	0.753	# 6
2D Human Pose Estimation	Human-Art	RTMPose-m	AP	0.355	# 10
2D Human Pose Estimation	Human-Art	RTMPose-m	AP (gt bbox)	0.728	# 8
2D Human Pose Estimation	Human-Art	RTMPose-s	AP	0.311	# 11
2D Human Pose Estimation	Human-Art	RTMPose-s	AP (gt bbox)	0.698	# 9
2D Human Pose Estimation	Human-Art	RTMPose-t	AP	0.249	# 12
2D Human Pose Estimation	Human-Art	RTMPose-t	AP (gt bbox)	0.655	# 10
Pose Estimation	OCHuman	RTMPose(RTMPose-l, GT bounding boxes)	Test AP	80.3	# 3
Pose Estimation	OCHuman	RTMPose(RTMPose-l, GT bounding boxes)	Validation AP	80.5	# 2

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/rtmpose-real-time-multi-person-pose/2d-human-pose-estimation-on-coco-wholebody-1)](https://paperswithcode.com/sota/2d-human-pose-estimation-on-coco-wholebody-1?p=rtmpose-real-time-multi-person-pose)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/rtmpose-real-time-multi-person-pose/pose-estimation-on-ochuman)](https://paperswithcode.com/sota/pose-estimation-on-ochuman?p=rtmpose-real-time-multi-person-pose)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/rtmpose-real-time-multi-person-pose/2d-human-pose-estimation-on-human-art)](https://paperswithcode.com/sota/2d-human-pose-estimation-on-human-art?p=rtmpose-real-time-multi-person-pose)`

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

13 Mar 2023 · Tao Jiang, Peng Lu, Li Zhang, Ningsheng Ma, Rui Han, Chengqi Lyu, Yining Li, Kai Chen ·

Recent studies on 2D pose estimation have achieved excellent performance on public benchmarks, yet its application in the industrial community still suffers from heavy model parameters and high latency. In order to bridge this gap, we empirically explore key factors in pose estimation including paradigm, model architecture, training strategy, and deployment, and present a high-performance real-time multi-person pose estimation framework, RTMPose, based on MMPose. Our RTMPose-m achieves 75.8% AP on COCO with 90+ FPS on an Intel i7-11700 CPU and 430+ FPS on an NVIDIA GTX 1660 Ti GPU, and RTMPose-l achieves 67.0% AP on COCO-WholeBody with 130+ FPS. To further evaluate RTMPose's capability in critical real-time applications, we also report the performance after deploying on the mobile device. Our RTMPose-s achieves 72.2% AP on COCO with 70+ FPS on a Snapdragon 865 chip, outperforming existing open-source libraries. Code and models are released at https://github.com/open-mmlab/mmpose/tree/1.x/projects/rtmpose.

PDF Abstract

Code

Add Remove Mark official

open-mmlab/mmpose official

4,971

Tasks

Add Remove

2D Human Pose Estimation

2D Pose Estimation

Multi-Person Pose Estimation

Pose Estimation

Datasets

MS COCO

MPII

CrowdPose

OCHuman

AP-10K

COCO-WholeBody AIC Human-Art

Results from the Paper

Edit

Ranked #3 on Pose Estimation on OCHuman (using extra training data)

Get a GitHub badge

Task	Dataset	Model	Metric Name	Metric Value	Global Rank	Benchmark
2D Human Pose Estimation	COCO-WholeBody	RTMPose	WB	65.3	# 3	Compare
			body	71.4	# 7	Compare
			foot	69.2	# 5	Compare
			face	88.9	# 1	Compare
			hand	59.0	# 4	Compare
2D Human Pose Estimation	Human-Art	RTMPose-l	AP	0.378	# 9	Compare
2D Human Pose Estimation	Human-Art	RTMPose-l	AP (gt bbox)	0.753	# 6	Compare
2D Human Pose Estimation	Human-Art	RTMPose-m	AP	0.355	# 10	Compare
2D Human Pose Estimation	Human-Art	RTMPose-m	AP (gt bbox)	0.728	# 8	Compare
2D Human Pose Estimation	Human-Art	RTMPose-s	AP	0.311	# 11	Compare
2D Human Pose Estimation	Human-Art	RTMPose-s	AP (gt bbox)	0.698	# 9	Compare
2D Human Pose Estimation	Human-Art	RTMPose-t	AP	0.249	# 12	Compare
2D Human Pose Estimation	Human-Art	RTMPose-t	AP (gt bbox)	0.655	# 10	Compare
Pose Estimation	OCHuman	RTMPose(RTMPose-l, GT bounding boxes)	Test AP	80.3	# 3	Compare
Pose Estimation	OCHuman	RTMPose(RTMPose-l, GT bounding boxes)	Validation AP	80.5	# 2	Compare

Methods

Add Remove

No methods listed for this paper. Add relevant methods here

Edit Social Preview

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

Code Edit Add Remove Mark official

Tasks Edit Add Remove

Datasets Edit

Results from the Paper Edit

Methods Edit Add Remove

Code

Add Remove Mark official

Tasks

Add Remove

Datasets

Results from the Paper

Edit

Methods

Add Remove