This repository contains an API for face detection and matching. The algorithm backend used for the computer vision is pluggable and customizable. Currently the following algorithms are supported:
More information on customizing and implementing new face detection algorithms can be found here.
- Install Docker, Docker-Compose and GNU Make.
- Clone this repo.
- Review the default configuration values in
- To run tests type
make testfrom within the top level directory.
- To initialize the database, type
make createdb. This only needs to be executed once.
- To start the server, run
Register your user by making a POST request to
Optionally retrieve a token by making a GET request to
/api/v1/tokenwith your username:password in the Authentication header. Access all other resources by passing your_token:any_value in the Authentication header (using Basic Auth) or by passing username:password for each request
Upload an image by making a POST request to
Process an image by making a POST request to
/api/v1/process_image. This will extract face vectors for the image and match it against all other images in the database. This is an asynchronous process.
Check the status of the image to see if it is finished processing by making a GET request to
/api/v1/process_image/<image_id>. Once the image is finished processing, it will be removed from the system's storage and only the face vectors and matches are persisted in the system's database.
See which other images are matches for a given image by making a GET request to