Inventory of open source face recognition projects
Data are as of July 31, 2023.
CompreFace
Exadel CompreFace is the leading free and open source facial recognition system.
This solution was only released on github in July 2020, and it looks promising. CompreFace makes our list of the best open source face recognition projects because it's one of the few self-hosted REST API face recognition solutions that can be launched with a single docker-compose command. REST API allows easy integration into your system without prior machine learning skills. Furthermore, it is scalable so that faces in multiple video streams can be recognized simultaneously.
CompreFace has a simple UI for managing user personas and face collections. It offers a choice between the two most popular face recognition methods: FaceNet (99.65% LFW accurate) and InsightFace (99.86% LFW accurate).
Exadel CompreFace is a free and open source face recognition GitHub project. Essentially, it's a docker-based application that can be used as a standalone server or deployed in the cloud. You don't need machine learning skills to set up and use CompreFace.
The system provides REST APIs for face recognition, face verification, face detection, landmark detection, mask detection, head pose detection, age and gender recognition. The solution also features a role management system (UI panel) that allows you to easily control who has access to your facial recognition services.
CompreFace is provided as a docker-compose configuration that supports different models working on CPU and GPU. Our solution is based on state-of-the-art methods and libraries such as FaceNet (99.65% LFW accuracy) and InsightFace (99.86% LFW accuracy).
Github 3.1k stars; it is still under active development;
Deepface
Deepface is a lightweight Python framework for face recognition and facial attribute analysis (age, gender, emotion, and race). It is a hybrid face recognition framework containing state-of-the-art models: VGG-Face, Google FaceNet, OpenFace, Facebook DeepFace, DeepID, ArcFace, Dlib and SFace.
Experiments have shown that humans are as high as 97.53% accurate on facial recognition tasks, and these models have met and exceeded this level of accuracy.
For developers who do not know Python, it is difficult to get started.
Github 7k stars; it is still under active development;
InsightFace
InsightFace is another open source Python library that uses one of the latest and most accurate face recognition methods for face detection (RetinaFace) and face recognition (SubCenter-ArcFace). The accuracy of this solution is very high - 99.86% on the LFW dataset. The only downside is that it's not easy to use.
Github 17k stars; it is still under active development;
Face Recognition
face_recognition is the most concise face recognition library in the world, you can use Python and command line tools to extract, recognize, and manipulate faces. Also, installation instructions for all major platforms and even docker images for quick setup are available on their github. Despite its popularity, the software has some drawbacks. The last release was in 2018 and there have been no major improvements since. It uses a rather outdated face recognition model with only 99.38% accuracy on LFW and no REST API. The recognition accuracy of children and Asian faces needs to be improved.
Github 48.9k stars; last updated in June 2022;
FaceNet
FaceNet is a popular open source Python library. The accuracy of this method is quite high - 99.65% on the LFW dataset, which is good but not the highest. The downside of this solution is that it doesn't have a REST API, and repositories are no longer supported.
This is the TensorFlow implementation of the face recognizer described in the paper "FaceNet.A Unified Embedding for Face Recognition and Clustering": TensorFlow implementation of the face recognizer described in . The project also uses ideas from the paper "Deep Face Recognition" by Oxford University's Visual Geometry Group.
Github 13.1k stars; last updated in April 2018;
face-api.js
face-api.js JavaScript API for face detection and face recognition using Tensorflow.js in browser and NodeJS.
Github 15.3k stars; last updated in April 2020;
OpenFace
Face recognition using deep neural networks.
OpenFace is a Python and Torch implementation of face recognition using deep neural networks, based on the CVPR 2015 paper FaceNet: Unified Embeddings for Face Recognition and Clustering by Google's Florian Schroff, Dmitry Kalenichenko, and James Philbin. Torch allows networks to be executed on the CPU or using CUDA.
Github 14.7k stars; last updated in November 2019;