Improving 2D Human Pose Estimation in Rare Camera Views with Synthetic Data

Miroslav Purkrabek, Jiri Matas
7/13/2023
cs.CV

No official code found. Best implementation: MiraPurkrabek/RePoGen (54 stars, GPL-3.0, last commit 2025-06-20).

Abstract

Methods and datasets for human pose estimation focus predominantly on side- and front-view scenarios. We overcome the limitation by leveraging synthetic data and introduce RePoGen (RarE POses GENerator), an SMPL-based method for generating synthetic humans with comprehensive control over pose and view. Experiments on top-view datasets and a new dataset of real images with diverse poses show that adding the RePoGen data to the COCO dataset outperforms previous approaches to top- and bottom-view pose estimation without harming performance on common views. An ablation study shows that anatomical plausibility, a property prior research focused on, is not a prerequisite for effective performance. The introduced dataset and the corresponding code are available on https://mirapurkrabek.github.io/RePoGen-paper/ .

View on arXivView PDF

Code Implementations(1)

MiraPurkrabek/RePoGenCommunity100%
GPL-3.0

[FG 24] The official repository of paper 'Improving 2D Human Pose Estimation across Unseen Camera Views with Synthetic Data'

546PythonJul 12, 20231 years agoGPL-3.0
datasethuman-pose-estimationpose-estimationresearchsynthetic-data

Similar Papers

PoseGen: Learning to Generate 3D Human Pose Dataset with NeRF

89% match1 repo102023

SynBody: Synthetic Dataset with Layered Human Models for 3D Human Perception and Modeling

88% match1 repo3082026

GTPT: Group-based Token Pruning Transformer for Efficient Human Pose Estimation

88% match1 repo192026

PoseGU: 3D Human Pose Estimation with Novel Human Pose Generator and Unbiased Learning

88% match2022

DeProPose: Deficiency-Proof 3D Human Pose Estimation via Adaptive Multi-View Fusion

87% match1 repo92025

Markerless Multi-view 3D Human Pose Estimation: a survey

87% match2024

Cite this paper

@article{purkrabek2023improving,
  title  = {Improving 2D Human Pose Estimation in Rare Camera Views with Synthetic Data},
  author = {Miroslav Purkrabek and Jiri Matas},
  year   = {2023},
  eprint = {2307.06737},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2307.06737v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2307.06737.svg)](https://codeofpaper.com/paper/2307.06737)

Discussion