It's using C++ language with FFMPEG, WebSocket, RTSP, OpenCV, Yolo, TensorFlow, SSDMobile, CNN, DNN, Dlib...
It could run on Embedded Computer such as Raspberry Pi, Orange Pi, Jetson Nano...
It can detect objects and recognize faces realtime and stream via RTSP.
Source code: https://github.com/tuaanpv/MBEDAI