The AI of the car uses sensor data to process all the data and classifies objects based on the size, shape and movement patterns. It can recognize surroundings from a 360 degree perspective by making predictions about vehicles, people and objects around it will move.
It can detect pedestrians, but as moving, column-shaped blurs of pixels, so it really cannot tell whether it's a rock or a crumpled piece of paper.
However it is programmed to determine certain patterns when a police officer has halted traffic or the car is being signaled to move forward.
It also recognizes cyclists as objects outlined in red and can slow down to let the cyclist enter into a lane.
Above images are provided by Chris Urmson who heads up Google's driverless car program.