The DOI matches its Crossref record; 32 of the 32 references it cites were found in a registry or an index.
Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields
Abstract
We present an approach to efficiently detect the 2D pose of multiple people in an image. The approach uses a nonparametric representation, which we refer to as Part Affinity Fields (PAFs), to learn to associate body parts with individuals in the image. The architecture encodes global context, allowing a greedy bottom-up parsing step that maintains high accuracy while achieving realtime performance, irrespective of the number of people in the image. The architecture is designed to jointly learn part locations and their association via two branches of the same sequential prediction process. Our method placed first in the inaugural COCO 2016 keypoints challenge, and significantly exceeds the previous state-of-the-art result on the MPII Multi-Person benchmark, both in performance and efficiency.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“For a stream of 100 simple dynamics images from the JAAD data set, this knob reduces the median image size by up to 40%.”
From Mez: An Adaptive Messaging System for Latency-Sensitive Multi-Camera Machine Vision at the IoT Edge (2021) · cites this paper as [19]
In the citing paper
“We note that for video frames with complex scene dynamics, the ratio FIVE Lat /ONE Lat is 5.6x for the JAAD dataset, and 8.4x for the DukeMTMC dataset.”
From Mez: An Adaptive Messaging System for Latency-Sensitive Multi-Camera Machine Vision at the IoT Edge (2021) · cites this paper as [19]
In the citing paper
“Although top-down methods still dominate the leaderboard of public benchmark datasets like MS COCO 10 , they are usually slower than bottom-up methods [112].”
From Empowering things with intelligence: a survey of the progress, challenges, and opportunities in artificial intelligence of things (2021) · cites this paper as [112]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 2×
“Insafutdinov et al. [15] built on [27] with a stronger part detectors based on ResNet [13] and image-dependent pairwise scores, and vastly improved the runtime with an incremental optimization approach, but the method still takes several minutes per image, with a limit of at most 150 part proposals.”
BackgroundMethodVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“During training, we initialize the 10 layers from the VGG-19 [32].”
MethodMicrosoft COCO: Common Objects in Context
Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., & Zitnick, C. L. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10602-1_48 · Cited by 45,460
(2016) · Cited by 31,590
The Hungarian method for the assignment problem
Kuhn, H. W. (1955) · Naval Research Logistics Quarterly · doi:10.1002/nav.3800020109 · Cited by 13,120
Stacked Hourglass Networks for Human Pose Estimation
Newell, A., Yang, K., & Deng, J. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46484-8_29 · Cited by 5,634 · cited 3×
“To obtain reliable local observations of body parts, Convolu-tional Neural Networks (CNNs) have been widely used, and have significantly boosted the accuracy on body pose estimation [23, 38, 24, 34, 35, 7, 36, 6].”
BackgroundMethodFields, J. (2001) · Cited by 4,044
DeepPose: Human Pose Estimation via Deep Neural Networks
Toshev, A., & Szegedy, C. (2014) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2014.214 · Cited by 3,395 · cited 1×
“To obtain reliable local observations of body parts, Convolu-tional Neural Networks (CNNs) have been widely used, and have significantly boosted the accuracy on body pose estimation [23, 38, 24, 34, 35, 7, 36, 6].”
MethodWei, S.-E., Ramakrishna, V., Kanade, T., & Sheikh, Y. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.511 · Cited by 2,966 · cited 7×
“…be capable of iteratively refining confidence maps by learning implicit spatial relationship between different body parts in the single person case [38]; we show the that sequential prediction architectures can also iteratively refine PAFs while capturing spatial relationship across different…”
BackgroundMethodInfluentialSelf-citation2D Human Pose Estimation: New Benchmark and State of the Art Analysis
Andriluka, M., Pishchulin, L., Gehler, P., & Schiele, B. (2014) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2014.471 · Cited by 2,699 · cited 1×
“We evaluate our method on two benchmarks for multi-person pose estimation: (1) the MPII human multi-person dataset [2] and (2) the MSCOCO 2016 keypoints challenge dataset [21].”
Method
Research works printed but not linked: 1
- 0, MSCOCO keypoint evaluation metric
Cited by
Sorted by citations
Items 1,001–1,016 of 1,016
Cognitive System Framework for Brain-Training Exercise Based on Human-Robot Interaction
Andriella, A., Torras, C., & Alenyà, G. (2020) · Cognitive Computation · doi:10.1007/s12559-019-09696-2 · Cited by 43
“One of the most promising alternatives we have evaluated is using OpenPose [47].”
MethodRey, V. F., Hevesi, P., Kovalenko, O., & Lukowicz, P. (2019) · Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers · doi:10.1145/3341162.3345590 · Cited by 43
“We used the OpenPose library [6, 7] to extract frame by frame the 2D poses of subjects.”
MethodDataset for human fall recognition in an uncontrolled environment
Guerrero, J. C. E., España, E. M., Añasco, M. M., & Lopera, J. E. P. (2022) · Data in Brief · doi:10.1016/j.dib.2022.108610 · Cited by 45
Miao, S., Chen, L., Hu, R., & Luo, Y. (2022) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3550331 · Cited by 36
“In skeleton-based HAR, the time series of 3D coordinates of multiple skeleton joints are estimated either from 2D images by pose estimation algorithms [6] or directly collected by depth cameras ( e.g., Kinetics) [58].”
MethodInterpersonal synchronization of spontaneously generated body movements
Koul, A., Ahmar, D., Iannetti, G. D., & Novembre, G. (2023) · iScience · doi:10.1016/j.isci.2023.106104 · Cited by 28
“Body-part-specific kinematics (specifically, movement velocities) were automatically extracted from video recordings using OpenPose, a human pose detection library relying on machine learning.(35) As an open-source and validated method, OpenPose achieves high accuracy in automatically extracting movement kinematics of…”
MethodBolt Loosening Detection Using Key-Point Detection Enhanced by Synthetic Datasets
Lu, Q., Jing, Y., & Zhao, X. (2023) · Applied Sciences · doi:10.3390/app13032020 · Cited by 28
Emotion Recognition from EEG and Facial Expressions: a Multimodal Approach
Chaparro, V., Gomez, A., Salgado, A., Quintero, O. L., Lopez, N., & Villa, L. F. (2018) · 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2018.8512407 · Cited by 27
Systemic Biases in Sign Language AI Research: A Deaf-Led Call to Reevaluate Research Agendas
Desai, A., De Meulder, M., Hochgesang, J. A., Kocab, A., & Lu, A. X. (2024) · Proceedings of the Language Resources and Evaluation Conference · doi:10.63317/4qt6t64rxwd2 · Cited by 25
Abbas, W., & Masip Rodo, D. (2019) · Sensors · doi:10.3390/s19153274 · Cited by 21
“For instance, Openpose can have a hard time identifying a pose it has not seen.”
MethodMujirishvili, T., Fedosov, A., Hashemifard, K., Climent-Pérez, P., & Florez-Revuelta, F. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642164 · Cited by 18
“New methods within this paradigm have shown substantial improvements and widespread applicability in various tasks such as action detection [89], pose estimation [12], anomaly detection [1], and fall detection [69].”
MethodFortes Rey, V., Garewal, K. K., & Lukowicz, P. (2021) · Applied Sciences · doi:10.3390/app11073094 · Cited by 17
“The application of standard tools (specifically Open Pose [21,41]) to identify humans in videos and extraction of their 2D poses.”
MethodKwon, H., Tong, C., Haresamudram, H., Gao, Y., Abowd, G. D., Lane, N. D., & Ploetz, T. (2020) · arXiv · doi:10.48550/arxiv.2006.05675 · Cited by 14
“The 2D pose estimator used in IMUTube, OpenPose, has previously known failure scenarios including partial detection of joints, swapping between left and right for rare poses, self-occlusion from the camera viewpoints, and partially visible bodies [10].”
MethodPerson Re-ID by Fusion of Video Silhouettes and Wearable Signals for Home Monitoring Applications
Masullo, A., Burghardt, T., Damen, D., Perrett, T., & Mirmehdi, M. (2020) · Sensors · doi:10.3390/s20092576 · Cited by 10
A Tensor‐based Regression Approach for Human Motion Prediction
Gril, L., Wedenig, P., Torkar, C., & Kleb, U. (2023) · Quality and Reliability Engineering International · doi:10.1002/qre.3153 · Cited by 7
Transferring Dense Pose to Proximal Animal Classes
Sanakoyeu, A., Khalidov, V., McCarthy, M. S., Vedaldi, A., & Neverova, N. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00528 · Cited by 13
“2D human pose recognition has flourished by the introduction of deep neural networks [56, 41, 10] trained on large manually-annotated datasets of images and videos such as COCO [30], MPII [4], Leeds Sports Pose Dataset (LSP) [20, 21], PennAction [61] and Posetrack [3].”
BackgroundPerson identity shift for privacy-preserving person re-identification
Dou, S., Jiang, X., Zhao, Q., Wang, Y., Li, D., & Zhao, C. (2026) · Science China Information Sciences · doi:10.1007/s11432-023-4583-x · Cited by 2
“For pose s , a trained multi-person key point estimation [49] is exploited to obtain the pedestrian posture.”
Metadata
- Authors
- Cao, Z., Simon, T., Wei, S.-E., & Sheikh, Y.
- Affiliations
- Carnegie Mellon University (US)
- Venue
- 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2017-07
- Pages
- 1302-1310
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W2559085405
- Cited by
- 7,539
- References
- 32
- Citations by year
- 2026: 384 · 2025: 774 · 2024: 968 · 2023: 1007 · 2022: 913 · 2021: 1016 · 2020: 1072 · 2019: 930 · 2018: 424 · 2017: 50
- FWCI
- 197.95
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- Funders
- Office of Naval Research
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Human Pose and Action Recognition
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Keywords
- multi-person pose estimation, part affinity fields, real-time pose estimation, bottom-up parsing, keypoint detection
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
