Skip to content

A containerized deployment of a Tensorflow-serving server and client container for classifying frames of video

License

Notifications You must be signed in to change notification settings

mit-ll/PSIAP-Video-Classifier-Deployment

PSIAP-Video-Classifier-Deployment

This sets up a pair of Docker container services for classifying footage with a tensorflow model

The two services are:

  • tf_server: A tensorflow-serving server which hosts Convolutional Neural Networks. The example model classifies footage as in-vehicle or out-of-vehicle. The services can be accessed from localhost at localhost:8501, or from within the docker-compose network at tf_server. See the tf_server README for more details
  • client: A python3 docker image which processes the videos contained within the subdirectory videos and generates a report in the subdirectory reports. It accomplishes this by using ffmpeg to extract keyframes, then submits those frames to tf_server for classification, and then generates a report based on the classifications. See the client README for more details

Windows Set-Up

  1. Create a Docker Hub Account
  2. Install Docker Desktop for Windows
  3. Add Docker to Windows Path, should be something like C:\Program Files\Docker\Docker\resources\bin
  4. Configure Docker to use proxy, if needed (note look for config.json in \Users\USERXXX.docker)
  5. Copy video files to directory: \deploy\videos
  6. Open Windows PowerShell
  7. Navigate to deploy directory
  8. Log in to the Docker registry, docker login
  9. Run Docker Compose to build service, docker-compose build
  10. docker-compose up
  11. docker-compose up -d

*nix instructions

  1. Install Docker for your platform, e.g. ubuntu
  2. Configure proxy if necessary
  3. Navigate into this directory in terminal
  4. Copy video files to directory ./videos
  5. docker-compose build
  6. docker-compose up

Authors

  • Jeffrey Liu (MITLL)
  • Chris Budny (MITLL)
  • Andrew Weinert (MITLL)

Acknowledgments

  • Gabriela Barrera (MITLL)
  • Dieter Schuldt (MITLL)
  • Steven Talpas (NJOHSP)
  • William Drew (NJOHSP)

Disclaimer

This work was performed under the following financial assistance award 70NANB17Hl69 from U.S. Department of Commerce, National Institute of Standards and Technology.

About

A containerized deployment of a Tensorflow-serving server and client container for classifying frames of video

Topics

Resources

License

Code of conduct

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •