Introduction
Classification answers "what is in this image?" with one label. Real scenes contain many things: a street photo has cars, people, bicycles, traffic lights. Object detection answers "what objects are here, and where is each one?" For every object it outputs a class and a bounding box (a rectangle around it), usually with a confidence score.
Think of a security guard looking at a camera feed and drawing a box around every person and every car, writing a label next to each. A detector does this for thousands of frames per minute. It powers self-driving cars, face detection in phone cameras, shelf monitoring in shops, and counting cells in microscope images.