Skip to content
Back to course

13.3 Object Detection

Object detection goes beyond classification by answering both 'what' and 'where'. For every object it finds, the model draws a bounding box and assigns a label and confidence. A single image can contain many objects of different classes, all detected at once.

faceobject
Detection outputs: each object gets a box, a class label and a confidence score.

Think of a traffic camera at a busy Addis Ababa junction. An object detector can count cars, buses, bajajs and pedestrians at the same time, each marked with its own box, which is impossible with plain classification.

Scenario

A warehouse needs to count and locate every parcel on a fast-moving belt in real time. Which approach fits best?

Check your understanding

1/4 · 40 XP

Object detection provides both the class and the location of each object.