CompreFace
A free, open-source face recognition server that runs in Docker, uses FaceNet and InsightFace, and supports CPU and GPU processing.
CompreFace is a self-hosted face recognition service for developers who want to add facial identification to an application without building or training their own machine learning system. It runs as a Docker-based server on your hardware or in a cloud deployment you manage. It's free and open source under the Apache 2.0 license.
Its REST API covers face detection, identifying people, and verifying whether faces match. Recognition can work with a single reference photo of a person. Plugins add facial landmark detection, age and gender estimation, mask detection, and head pose detection, so applications can extract information beyond identity through the same service.
CompreFace uses FaceNet and InsightFace and supports models that run on either a CPU or a GPU. The stated hardware requirement is an x86 processor with AVX support. It runs through Docker on Linux, macOS, and Windows, making it an option for teams that want to host face processing on their own infrastructure. An on-premises deployment puts the recognition service on a server they control; a cloud deployment runs it on cloud infrastructure.
A browser-based administration panel manages user roles and access to face recognition services. That matters when multiple users need different permissions on a shared server. JavaScript, Python, and .NET SDKs provide language-specific ways to connect applications to the API.
