Video Query Algorithms

Home page for Video Query project: Video Query Home

This respository provides the algorithms and code for:

• building a database of embedded DNN features from videos,
• brokering requests from the Video Query API to perform searches of the database,
• searching the database for video clips similar to a provided reference, and
• updating the weighted criteria for selecting matches.

The code and documentation for the Client and API are at

• Video Query Client
• Video Query Api

For further information about the Video Query project, please go to Video Query Home.

Setup

After cloning this repository, you will be required to set the following Environment Variables to run this project

• API_CLIENT_USERNAME = a valid username for the API
• API_CLIENT_PASSWORD = password for user API_CLIENT_USERNAME
• TSN_ROOT
• TSN_ENVIRON
• BROKER_THREADING = True or False
  • True: each time the broker queries the API for new jobs, a new broker thread is also launched. this is the default.
  • False: broker only checks once for jobs, then stops after executing any waiting jobs. Subsequent jobs won't be run until broker.py is manually run again, once per job. This setting is for debugging.
• COMPUTE_EPS = a small number, e.g. 0.000003
• RANDOM_SEED = random integer for seeding the python random numbers. Setting this enables checking for reproducibility. Example: "export RANDOM_SEED=73459912436" will enable a reproducible set of code executions in each call of compute_matches.py. Setting RANDOM_SEED=None will result in setting a random seed based on system time, which is more appropriate when not debugging or testing code.

One way to set these is to execute

bash set_environ.sh in Linux or
source set_environ.sh in MacOS,

using a set_environ.sh file in your local video-query-algorithms main directory. Here is an example of a set_environ.sh file used for development:

#!/usr/bin/env bash
export TSN_ROOT='/data/torres/temporal-segment-networks'
export TSN_ENVIRON='TSN'
export API_CLIENT_USERNAME='your_username'
export API_CLIENT_PASSWORD='your_password'
export BROKER_THREADING='True'
export COMPUTE_EPS=.000003
export RANDOM_SEED=73459912436abcd

Wiki

For detailed instruction on how to utilize this repository, please have a look at our Wiki

• Algorithms
  • Pipeline
  • Compute video features
  • Database
  • Conda
  • Broker

Contributing

Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.

Credits

Include a citation for a paper when we publish one. For now: If you find this repository useful, please cite https://github.com/PARC-projects/video-query-home.

Project team

Software development:

• Frank Torres
• Chad Ramos

Algorithm development by Frank Torres, Matthew Shreve, Gaurang Ganguli and Hoda Eldardiry.

License

Copyright (C) 2018 Palo Alto Research Center, Inc.

This program is free software for non-commercial uses: you can redistribute it and/or modify it under the terms of the Aladdin Free Public License - see the LICENSE.md file for details.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

If this license does not meet your needs, please submit an Issue on github with your contact information and institution, or email engage@parc.com, so we can discuss how to meet your needs.