DataIker Facial Recognition

I created this project during my work experience at data science company DataIku. It consists of a Raspberry Pi that uses OpenCV and an onboard camera to identify human faces that it can see. In its 'learning' mode, it will store each captured image of a face on an Amazon S3 bucket. These images can then be labelled using a webapp that I created. I created a machine learning model on the company's bespoke software, DSS, which can use these labelled images to learn to distinguish the different employees of DataIku (or 'DataIkers') by name using only their face.

If the device is not in 'learning' mode, it will instead use an API endpoint on DSS to attempt to identify any office member that walks past the camera, and greet them by name on the LCD screen attached to the Pi.