CompreFace 0.5.0 #504

pospielov · 2021-04-25T09:46:07Z

pospielov
Apr 25, 2021
Maintainer

New services. We added face detection and face verification services to cover more business cases.
Face plugins. Enjoy age and gender recognition features and receive face landmarks and calculated features.
Added scalability. Scale your face recognition system on a dozen servers to increase your system throughput.
Clone face recognition service. Duplicate your face recognition collection and test new data on it. Your application won’t be affected!
Update user info and password. We provided a better user experience by adding the ability to update user’s profiles.
Support of Insightface face recognition models. Insightface is a state-of-the-art library with a variety of models for different devices and cases.
GPU support. Use all power of your GPU accelerator to enjoy the increased speed of the system.

You need to place new docker-compose.yml and .env files into the current CompreFace config directory.
We updated REST API for more clear service distinction. Face collections now have a "/recognition" context. This means you'll need to update your application logic to use new REST endpoints.
All Face Collections were migrated to Face Recognition Services automatically, so no actions from users are required.
We updated the formula for calculating the similarity between faces. So if before the recommended threshold was 0.7, now it's 0.5 for every model. You will need to adjust the threshold in your application logic.

Download CompreFace_0.5.0.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/0.5.0/CompreFace_0.5.0.zip' && unzip tmp.zip && rm tmp.zip
To start CompreFace run:
docker-compose up -d
Open in your browser: http://localhost:8000/login
To stop properly run:
docker-compose stop
To start again run:
docker-compose start

More detail install guide(and Windows instructions) is on our readme page.

This discussion was created from the release CompreFace 0.5.0.