mlsanityData to 14 Oct 2026

HSTGFormer: Hyper Spatial-Temporal Graph Transformer for 3D Human Pose Estimation

arXiv 2608.12187 ↗First posted 12 Aug 2026cs.CV cs.AIcited by 0

On the map: LLM reasoning → Spatial and embodied reasoning

Abstract (arXiv)

Transformer-based methods have achieved strong performance in monocular 3D human pose estimation, but most existing approaches organise spatial and temporal reasoning as separate stages, which may weaken unified spatial-temporal interdependencies inherent in human motion and compress frame-level structural information before temporal modelling. In this paper, we propose HSTGFormer, a graph-enhanced Transformer framework that reformulates spatial-temporal reasoning as localised coupled graph aggregation over joint-time nodes. Specifically, HSTGFormer introduces a Hyper Spatial-Temporal Graph (HSTG), which decomposes global spatial-temporal reasoning into local spatial-temporal receptive fields around individual joint-time nodes by extending per-frame skeleton graphs into temporal neighbourhoods, thereby enabling structure-aware coupled reasoning while preserving local structural motion information. It further incorporates an Adaptive Dual-Scale Temporal Graph (ADSTG) to capture joint-specific temporal dependencies over complementary short- and long-range windows. A lightweight node-wise fusion module further adaptively integrates the two graph representations for each joint-time node. Experiments on Human3.6M and MPI-INF-3DHP show that HSTGFormer achieves strong accuracy with high computational efficiency.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 34 of its 42 references that are in the map. First posted on arXiv 12 Aug 2026.

Builds on

34 of 42 references.

References · 34

  1. 2017Attention Is All You Needcited by 60,361
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  4. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  5. 2017Graph Attention Networkscited by 8,923
  6. 2016Layer Normalizationcited by 5,669
  7. 2016Stacked Hourglass Networks for Human Pose Estimationcited by 915
  8. 2016Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervisioncited by 494
  9. 2017A simple yet effective baseline for 3d human pose estimationcited by 401
  10. 20213D Human Pose Estimation with Spatial and Temporal Transformerscited by 272
  11. 2022MotionBERT: A Unified Perspective on Learning Human Motion Representationscited by 193
  12. 2022MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Videocited by 159
22 more
  1. 2021MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimationcited by 151
  2. 2023FourierGNN: Rethinking Multivariate Time Series Forecasting from a Pure Graph Perspectivecited by 127
  3. 2022DiffPose: Toward More Reliable 3D Pose Estimationcited by 100
  4. 2023PoseFormerV2: Exploring Frequency Domain for Efficient and Robust 3D Human Pose Estimationcited by 92
  5. 2022P-STMO: Pre-Trained Spatial Temporal Many-to-One Model for 3D Human Pose Estimationcited by 80
  6. 2023MotionAGFormer: Enhancing 3D Human Pose Estimation with a Transformer-GCNFormer Networkcited by 64
  7. 2023GLA-GCN: Global-local Adaptive Graph Convolutional Network for 3D Human Pose Estimation from Monocular Videocited by 47
  8. 2024KTPFormer: Kinematics and Trajectory Prior Knowledge-Enhanced Transformer for 3D Human Pose Estimationcited by 34
  9. 2023SportsPose -- A Dynamic 3D sports pose datasetcited by 26
  10. 2025TCPFormer: Learning Temporal Correlation with Implicit Pose Proxy for 3D Human Pose Estimationcited by 14
  11. 2025Unified Spatial-Temporal Edge-Enhanced Graph Networks for Pedestrian Trajectory Predictioncited by 13
  12. 2022Multiclass-SGCN: Sparse Graph-based Trajectory Prediction with Agent Class Embeddingcited by 9
  13. 2025BP-SGCN: Behavioral Pseudo-Label Informed Sparse Graph Convolution Network for Pedestrian and Heterogeneous Trajectory Predictioncited by 9
  14. 20213D Pose Estimation and Future Motion Prediction from 2D Imagescited by 8
  15. 2025Over-squashing in Spatiotemporal Graph Neural Networkscited by 6
  16. 2025AthleticsPose: Authentic Sports Motion Dataset on Athletic Field and Evaluation of Monocular 3D Pose Estimation Abilitycited by 6
  17. 2025H2OT: Hierarchical Hourglass Tokenizer for Efficient Video Pose Transformerscited by 5
  18. 2024From Category to Scenery: An End-to-End Framework for Multi-Person Human-Object Interaction Recognition in Videoscited by 4
  19. 2025ViTE: Virtual Graph Trajectory Expert Router for Pedestrian Trajectory Predictioncited by 4
  20. 2025Geometric Visual Fusion Graph Neural Networks for Multi-Person Human-Object Interaction Recognition in Videoscited by 3
  21. 2026Motion-Adaptive Multi-Scale Temporal Modelling with Skeleton-Constrained Spatial Graphs for Efficient 3D Human Pose Estimationcited by 2
  22. 2026ART: Adaptive Relational Transformer for Pedestrian Trajectory Prediction with Temporal-Aware Relationscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026NCSTR: Node-Centric Decoupled Spatio-Temporal Reasoning for Video-based Human Pose Estimation
  2. 2025PRGCN: A Graph Memory Network for Cross-Sequence Pattern Reuse in 3D Human Pose Estimation
  3. 2026HYPERPOSE: Hyperbolic Kinematic Phase-Space Attention for 3D Human Pose Estimation
  4. 2025Efficient Diffusion-Based 3D Human Pose Estimation with Hierarchical Temporal Pruning
  5. 2023Multiple View Geometry Transformers for 3D Human Pose Estimation
  6. 2026Hypergraph-State Collaborative Reasoning for Multi-Object Tracking
  7. 2025HDiffTG: A Lightweight Hybrid Diffusion-Transformer-GCN Architecture for 3D Human Pose Estimation
  8. 2025SCJD: Sparse Correlation and Joint Distillation for Efficient 3D Human Pose Estimation