# Automatic infant 2D pose estimation from videos: comparing seven deep neural network methods

arXiv: https://arxiv.org/abs/2406.17382 · HTML: https://mlsanity.com/paper/2406.17382/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 70 references that are in the map.
- First posted on arXiv 25 Jun 2024.

## Builds on (most cited first, 18 of 18)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Expressive Body Capture: 3D Hands, Face, and Body from a Single Image](https://mlsanity.com/paper/1904.05866/) (2019)
- [Deep High-Resolution Representation Learning for Human Pose Estimation](https://mlsanity.com/paper/1902.09212/) (2019)
- [OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields](https://mlsanity.com/paper/1812.08008/) (2018)
- [MediaPipe: A Framework for Building Perception Pipelines](https://mlsanity.com/paper/1906.08172/) (2019)
- [Simple Baselines for Human Pose Estimation and Tracking](https://mlsanity.com/paper/1804.06208/) (2018)
- [ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation](https://mlsanity.com/paper/2204.12484/) (2022)
- [Humans in 4D: Reconstructing and Tracking Humans with Transformers](https://mlsanity.com/paper/2305.20091/) (2023)
- [Deep Learning-Based Human Pose Estimation: A Survey](https://mlsanity.com/paper/2012.13392/) (2020)
- [Markerless tracking of user-defined features with deep learning](https://mlsanity.com/paper/1804.03142/) (2018)
- [AlphaPose: Whole-Body Regional Multi-Person Pose Estimation and Tracking in Real-Time](https://mlsanity.com/paper/2211.03375/) (2022)
- [BlazePose: On-device Real-time Body Pose tracking](https://mlsanity.com/paper/2006.10204/) (2020)
- [DeeperCut: A Deeper, Stronger, and Faster Multi-Person Pose Estimation Model](https://mlsanity.com/paper/1605.03170/) (2016)
- [Benchmarking and Error Diagnosis in Multi-Instance Pose Estimation](https://mlsanity.com/paper/1707.05388/) (2017)
- [Invariant Representation Learning for Infant Pose Estimation with Small Data](https://mlsanity.com/paper/2010.06100/) (2020)
- [Movement science needs different pose tracking algorithms](https://mlsanity.com/paper/1907.10226/) (2019)
- [Automatic Posture and Movement Tracking of Infants with Wearable Movement Sensors](https://mlsanity.com/paper/1909.09823/) (2019)
- [Towards human-level performance on automatic pose estimation of infant spontaneous movements](https://mlsanity.com/paper/2010.05949/) (2020)

## Built on by (most cited first, 6 of 6)

- [Comparison of marker-less 2D image-based methods for infant pose estimation](https://mlsanity.com/paper/2410.04980/) (2024)
- [Markerless Motion Capture for Biomechanical Whole-Body Kinematic Estimation in Infants](https://mlsanity.com/paper/2605.17120/) (2026)
- [The Blind Spot in 2D Infants' Pose Estimation:Robust Learning from Noisy Annotations](https://mlsanity.com/paper/2609.04009/) (2026)
- [BBoxMaskPose v2: Expanding Mutual Conditioning to 3D](https://mlsanity.com/paper/2601.15200/) (2026)
- [Synthetic-Child: An AIGC-Based Synthetic Data Pipeline for Privacy-Preserving Child Posture Estimation](https://mlsanity.com/paper/2603.02598/) (2026)
- [Cross-Model Distillation of a Human-Pose Foundation Model from Unannotated Infant Video for Markerless 3D Pose Estimation](https://mlsanity.com/paper/2609.01840/) (2026)

## Abstract (arXiv)

Automatic markerless estimation of infant posture and motion from ordinary videos carries great potential for movement studies "in the wild", facilitating understanding of motor development and massively increasing the chances of early diagnosis of disorders. There is rapid development of human pose estimation methods in computer vision thanks to advances in deep learning and machine learning. However, these methods are trained on datasets that feature adults in different contexts. This work tests and compares seven popular methods (AlphaPose, DeepLabCut/DeeperCut, Detectron2, HRNet, MediaPipe/BlazePose, OpenPose, and ViTPose) on videos of infants in supine position and in more complex settings. Surprisingly, all methods except DeepLabCut and MediaPipe have competitive performance without additional finetuning, with ViTPose performing best. Next to standard performance metrics (average precision and recall), we introduce errors expressed in the neck-mid-hip (torso length) ratio and additionally study missed and redundant detections, and the reliability of the internal confidence ratings of the different methods, which are relevant for downstream tasks. Among the networks with competitive performance, only AlphaPose could run close to real time (27 fps) on our machine. We provide documented Docker containers or instructions for all the methods we used, our analysis scripts, and the processed data at https://hub.docker.com/u/humanoidsctu and https://osf.io/x465b/.