cmusatyalab/openface

OpenFace is a free, open-source library for face recognition, the technology that identifies or verifies a person's identity from a photo or video frame. It was developed by researchers at Carnegie Mellon University with support from the National Science Foundation, Intel, Google, Vodafone, and NVIDIA. The library uses deep neural networks (a type of artificial intelligence that learns patterns from large amounts of data) to process face images and produce numerical "representations", compact descriptions of a face that can be compared against other faces to check for a match. The codebase includes a Python library for core functionality, scripts for training new neural network models, tools to evaluate accuracy, and several ready-to-run demos. Those demos cover comparing two images side by side, running a real-time web demo, classifying faces using a trained model, and even identifying faces from a live webcam stream. The repository also includes batch processing tools to generate representations from large collections of images at once. It is licensed under the Apache 2.0 License, meaning it can be used and modified freely for most purposes including commercial ones. The underlying research is documented in a Carnegie Mellon technical report titled "OpenFace: A general-purpose face recognition library with mobile applications," published in 2016.