SOTAVerified

3D Human Pose Estimation

3D Human Pose Estimation is a computer vision task that involves estimating the 3D positions and orientations of body joints and bones from 2D images or videos. The goal is to reconstruct the 3D pose of a person in real-time, which can be used in a variety of applications, such as virtual reality, human-computer interaction, and motion analysis.

Papers

Showing 1–25 of 665 papers

TitleStatusHype
BlazePose: On-device Real-time Body Pose trackingCode4
TRAM: Global Trajectory and Motion of 3D Humans from in-the-wild VideosCode4
PIFuHD: Multi-Level Pixel-Aligned Implicit Function for High-Resolution 3D Human DigitizationCode3
Humans in 4D: Reconstructing and Tracking Humans with TransformersCode3
TRACE: 5D Temporal Regression of Avatars with Dynamic Cameras in 3D EnvironmentsCode3
HybrIK-X: Hybrid Analytical-Neural Inverse Kinematics for Whole-body Mesh RecoveryCode3
Multi-HMR: Multi-Person Whole-Body Human Mesh Recovery in a Single ShotCode3
SMPLer-X: Scaling Up Expressive Human Pose and Shape EstimationCode3
WHAM: Reconstructing World-grounded Humans with Accurate 3D MotionCode3
MotionBERT: A Unified Perspective on Learning Human Motion RepresentationsCode3
TokenHMR: Advancing Human Mesh Recovery with a Tokenized Pose RepresentationCode3
PromptHMR: Promptable Human Mesh RecoveryCode3
MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in VideoCode2
A Survey on 3D Egocentric Human Pose EstimationCode2
MotionAGFormer: Enhancing 3D Human Pose Estimation with a Transformer-GCNFormer NetworkCode2
KTPFormer: Kinematics and Trajectory Prior Knowledge-Enhanced Transformer for 3D Human Pose EstimationCode2
3D Human Mesh Estimation from Virtual MarkersCode2
ICON: Implicit Clothed humans Obtained from NormalsCode2
Learning 3D Human Pose Estimation from Dozens of Datasets using a Geometry-Aware Autoencoder to Bridge Between Skeleton FormatsCode2
Human-Art: A Versatile Human-Centric Dataset Bridging Natural and Artificial ScenesCode2
Hulk: A Universal Knowledge Translator for Human-Centric TasksCode2
H3WB: Human3.6M 3D WholeBody Dataset and BenchmarkCode2
FrankMocap: A Monocular 3D Whole-Body Pose Estimation System via Regression and IntegrationCode2
AthletePose3D: A Benchmark Dataset for 3D Human Pose Estimation and Kinematic Validation in Athletic MovementsCode2
Hourglass Tokenizer for Efficient Transformer-Based 3D Human Pose EstimationCode2
Show:102550
← PrevPage 1 of 27Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1Simple-baselinePA-MPJPE157—Unverified
2HMRMPJPE130—Unverified
3BMPMPVPE119.3—Unverified
4SPINMPVPE116.4—Unverified
5Wenshuo et a;.MPVPE112.6—Unverified
6TCMR (T=16 w/o 3DPW)MPVPE111.5—Unverified
7CHOMPMPVPE110.1—Unverified
8PC-HMRMPVPE108.6—Unverified
93DCrowdNetMPVPE108.5—Unverified
10SMPLifyPA-MPJPE106.8—Unverified
#ModelMetricClaimedVerifiedStatus
1VNect (Augm.)MPJPE124.7—Unverified
2HMRMPJPE124.2—Unverified
3Single-Shot Multi-PersonMPJPE122.2—Unverified
4MehtaMPJPE117.6—Unverified
5PONetMPJPE115—Unverified
6Pose Consensus (monocular)MPJPE112.1—Unverified
7GeoRep (fully-supervised)MPJPE110.8—Unverified
8XFormer (HRNet)MPJPE109.8—Unverified
9EpipolarPose (fully-supervised)MPJPE108.99—Unverified
10SPINMPJPE105.2—Unverified