Based on a fast neural network architecture, our race recognition module can be easily integrated into applications that require accurate estimation of the people race from facial images. It can operate on embedded hardware, on-premise servers or can be deployed as cloud API. The recognized races are: White, Hispanic, Black, Asian and East Indian.
There are lots of industries that could benefit from our face recognition module, including security, marketing and advertising. Typical applications, to name a few are:
Intelligent Video Surveillance
Tagging of Video and Images
The race classifier is a neural network model in TensorFlow format. It is is based on the MobileNet neural network architecture, which allows running in real time on CPU. It takes 35 milliseconds on Intel i5 CPU for one face classification.