Skip to main content

Web-browser demo for openpifpaf.

Project description



Links: main repository, 2021 paper and CVPR2019 paper.
Live demo:
Serverless live demo:

Use Locally

Run the full pipeline (neural network and decoder) in Python and visualize the output in the browser:

pip3 install openpifpafwebdemo
python3 -m openpifpafwebdemo.server

Troubleshooting: Make sure you are using Python3 and have the latest pip and setuptools with pip install --upgrade pip setuptools. Do not clone this repository. Make sure there is no folder named openpifpafwebdemo in your current directory.


example image


For every release, this repository builds a new docker image and uploads it to Docker Hub: svenkreiss/openpifpafwebdemo. You can test it with:

docker run -d -p 5000:80 svenkreiss/openpifpafwebdemo

This docker images also powers More info on building the docker image and how to deploy it to Kubernetes is in docker/


Example using cURL:

curl --data-binary @docs/me_nyc_square_500.jpeg http://localhost:5000/v1/human-poses

which produces:

[{"coordinates": [[0.588631883263588, 0.41628291457891464, 3.5567557387194797], [0.621234196703881, 0.36160339042544365, 3.524825929280572], [0.546875, 0.375, 3.744302039019678], [0.6724068783223629, 0.44710323959589005, 3.459401266884038], [0.494683139026165, 0.4611567258834839, 3.5954212359489217], [0.78733691573143, 0.8311769068241119, 2.1321910543190827], [0.3859005756676197, 0.8252473473548889, 2.158424186304439], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], "score": 0.26909651332876167}]

Keep-alive connection test:

curl -v --data-binary @docs/me_nyc_square_500.jpeg http://localhost:5000/v1/human-poses --next --data-binary @docs/me_nyc_square_500.jpeg http://localhost:5000/v1/human-poses 2>&1 | grep '#0'
* Connected to localhost ( port 5000 (#0)
* Connection #0 to host localhost left intact
* Re-using existing connection! (#0) with host localhost
* Connected to localhost ( port 5000 (#0)
* Connection #0 to host localhost left intact


To install from source and set up for development use pip install --editable ".[test]", install the frontend dependencies with npm install and then create the frontend JavaScript code with npm run build. For continuous rebuilds of the js package, use npm run watch.

Run the server process with --debug to get salted version numbers to break the browser cache for static assets and autoreload when source files change.


  title = {{OpenPifPaf: Composite Fields for Semantic Keypoint Detection and Spatio-Temporal Association}},
  author = {Sven Kreiss and Lorenzo Bertoni and Alexandre Alahi},
  journal = {arXiv preprint arXiv:2103.02440},
  month = {March},
  year = {2021}

  author = {Kreiss, Sven and Bertoni, Lorenzo and Alahi, Alexandre},
  title = {PifPaf: Composite Fields for Human Pose Estimation},
  booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  month = {June},
  year = {2019}

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for openpifpafwebdemo, version 0.12.6.post2
Filename, size File type Python version Upload date Hashes
Filename, size openpifpafwebdemo-0.12.6.post2.tar.gz (506.1 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page