# facebookresearch/videopose3d

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/facebookresearch-videopose3d).**

3,986 stars · 782 forks · Python · other · archived

## Links

- GitHub: https://github.com/facebookresearch/VideoPose3D
- awesome-repositories: https://awesome-repositories.com/repository/facebookresearch-videopose3d.md

## Description

VideoPose3D is a machine learning framework designed for 3D human pose estimation. It functions as a motion reconstruction tool that predicts 3D joint positions from 2D video sequences using a temporal convolutional network to process body movement over time.

The project includes a semi-supervised learning pipeline that improves pose accuracy by combining labeled datasets with unlabeled video data and projection consistency loss. It also features a video pose visualizer capable of rendering 3D skeleton reconstructions and 2D keypoints as overlays on original footage.

The framework covers the full lifecycle of motion analysis, including 3D pose model training, biometric motion validation against ground truth data, and the generation of visual predictions in GIF or MP4 formats.

## Tags

### Artificial Intelligence & ML

- [3D Pose Estimation](https://awesome-repositories.com/f/artificial-intelligence-ml/3d-pose-estimation.md) — Provides a framework for converting 2D pose detections from video into three-dimensional spatial coordinates. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/videopose3d@main/README.md))
- [Motion Reconstruction](https://awesome-repositories.com/f/artificial-intelligence-ml/motion-reconstruction.md) — Implements algorithms to convert visual tracking data into 3D skeletal motion coordinates.
- [Semi-supervised Learning Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/semi-supervised-learning-pipelines.md) — Features a training pipeline that processes labeled and unlabeled video data simultaneously using projection consistency loss.
- [Temporal Convolutional Networks](https://awesome-repositories.com/f/artificial-intelligence-ml/time-series-forecasting/temporal-convolutional-networks.md) — Employs a temporal convolutional network architecture to reconstruct 3D body movement over time.
- [Temporal Pose Analysis](https://awesome-repositories.com/f/artificial-intelligence-ml/time-series-forecasting/temporal-convolutional-networks/temporal-pose-analysis.md) — Uses temporal convolutions to estimate human motion over time from video sequences.
- [Custom Video Inference](https://awesome-repositories.com/f/artificial-intelligence-ml/3d-pose-estimation/custom-video-inference.md) — Enables the analysis of original video files to calculate 3D human poses in real-world environments. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/videopose3d@main/README.md))
- [3D Pose Reconstruction](https://awesome-repositories.com/f/artificial-intelligence-ml/3d-pose-reconstruction.md) — Visualizes the mapping of 2D tracking data into three-dimensional skeletal space for validation. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/VideoPose3D@main/DOCUMENTATION.md#documentation))
- [Pose Estimation Validation](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-evaluation-and-validation/pose-estimation-validation.md) — Includes routines for verifying the precision of pose reconstructions against ground truth biometric data.
- [3D Pose Model Training](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training/pose-model-architectures/3d-pose-model-training.md) — Develops models that estimate 3D joint positions from 2D detections using temporal convolutions. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/VideoPose3D@main/DOCUMENTATION.md#documentation))
- [Model Validation Tools](https://awesome-repositories.com/f/artificial-intelligence-ml/model-validation-tools.md) — Includes utilities to verify reconstruction precision against specific subjects or actions using saved checkpoints. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/VideoPose3D@main/DOCUMENTATION.md#documentation))

### Graphics & Multimedia

- [Pose Visualizations](https://awesome-repositories.com/f/graphics-multimedia/pose-visualizations.md) — Provides a visualizer for rendering 3D skeleton reconstructions and 2D keypoints as video overlays.
- [Visual Overlays](https://awesome-repositories.com/f/graphics-multimedia/visual-overlays.md) — Generates MP4 or GIF overlays of 3D model predictions onto original video footage. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/videopose3d@main/README.md))