https://github.com/oveddan/runwayml-gazecapture

This repository brings the pre-trained model from Eye Tracking for Everyone into python and RunwayML, allowing it to be easily Dockerized and deployed
https://github.com/oveddan/runwayml-gazecapture

Last synced: 15 days ago
JSON representation

This repository brings the pre-trained model from Eye Tracking for Everyone into python and RunwayML, allowing it to be easily Dockerized and deployed

• Host: GitHub
• URL: https://github.com/oveddan/runwayml-gazecapture
• Owner: oveddan
• Created: 2018-05-02T14:00:07.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2018-05-06T17:02:24.000Z (over 6 years ago)
• Last Synced: 2024-10-11T01:23:54.263Z (about 1 month ago)
• Language: Python
• Homepage:
• Size: 22.5 KB
• Stars: 30
• Watchers: 2
• Forks: 6
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# GazeCapture RunwayML

This repository brings the pre-trained model from [Eye Tracking for Everyone](https://github.com/CSAILVision/GazeCapture)

into python and RunwayML.

It allows for a server to be spun up in a docker container that performs real-

time gaze estimation from a video stream.  It works with any webcam.

I have used this model successfully in my project [Presence](http://www.danioved.com/portfolio/presence/) - a kinetic sculpture that reacts to a users gaze.

### The server:

Install Docker.

Build the docker container in a tag:

    docker build . -t gaze

Launch the container:

    ./start_container.sh

### The client:

The client opens a webcam video feed and sends it in a stream to the server, getting gaze positions back.

Install dependencies:

    pip install opencv-python zmq

        

Launch the client:

    python test_client.py