The Google Vision APIs provide two main areas of functionality.
First is Face Tracking -- not to be confused with Facial Recognition -- which gives your apps the ability to detect faces, and landmarks on faces. This is useful for, for example -- writing a camera app that only takes a picture when everyone is smiling, and nobody is blinking, or for fun apps where you can superimpose hats or moustaches on people in the camera preview window.
Second is recognizing visual codes such as bar codes or QR codes, and making it easy for developers to build apps with them.
This Dev Byte covers the bar codes and QR codes. A separate byte will cover face detection