https://github.com/hayden-donnelly/pose-estimation
Demo of movenet pose estimation. Can be used to create videos with overlayed pose estimations for up to six people.
https://github.com/hayden-donnelly/pose-estimation
computer-vision machine-learning opencv pose-estimation
Last synced: 9 months ago
JSON representation
Demo of movenet pose estimation. Can be used to create videos with overlayed pose estimations for up to six people.
- Host: GitHub
- URL: https://github.com/hayden-donnelly/pose-estimation
- Owner: hayden-donnelly
- License: mit
- Created: 2023-03-29T01:24:56.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2023-06-22T22:21:55.000Z (about 3 years ago)
- Last Synced: 2024-11-19T01:52:12.011Z (over 1 year ago)
- Topics: computer-vision, machine-learning, opencv, pose-estimation
- Language: Python
- Homepage:
- Size: 522 KB
- Stars: 1
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# pose-estimation
Demo of movenet pose estimation. Can be used to create videos with overlayed pose estimations for up to six people.
[Singlepose Example](https://youtu.be/N1KlFnjoEtg), [Multipose Example](https://youtu.be/XxJNebC_oqc)
## Usage
Run ``movenet.py`` with ``--input_path`` set to the input video path. By default the output will be saved to ``./data``.
Example:
```
python movenet.py --path ../data/input/pose_estimation_video.mp4
```
Full list of parameters:
- ``--input_path``: Path to input video file. Default: ``None``
- ``--output_path``: Path to output directroy. Default: ``./data``
- ``--confidence_threshold``: Confidence threshold for keypoints. Default: ``0.1``
## Docker Environment
Building image:
```
docker-compose build
```
Starting container/environment:
```
docker-compose up -d
```
Opening a shell in container:
```
docker-compose exec pose-estimation bash
```
Instead of opening a shell, you can also go to http://localhost:8888/ to access a Jupyter Lab instance running inside the container.
Stopping container/environment:
```
docker-compose down
```
## Example Output
