Index Ai Academy
Glossary · Phase 5 · Transformers

Bounding box

Also written: bounding boxes

The smallest upright rectangle that holds an object, written as two corners or as a center with a width and height.

A detection has three parts: a bounding box, the smallest upright rectangle that holds the object; a class, such as dog; and a score from 0 to 1 saying how confident the detector is.

Try it now

Learn it properly

Object detection: boxes, overlap and suppression · Every object and where it is: boxes in two formats, intersection over union, why one object gets many boxes, non-maximum suppression and its threshold, and grading with average precision. About 45 minutes.

Other terms from the same lesson

  • intersection over union: How well two boxes or regions agree: the area they share divided by the area either covers. 1 is identical, 0 is no overlap.
  • mean average precision: A detector's average precision, averaged over all its classes, and often over several overlap thresholds too.
  • non-maximum suppression: Cutting a cluster of overlapping detections down to one: keep the highest-scoring box, drop those that overlap it by more than a threshold, and repeat.
  • object detection: Finding every object in an image and where it is: a box, a class and a confidence score for each.
Keep learning

More in plain English.

1286 terms in the glossary, each with the lesson that teaches it.

1286 terms

The glossary

Every word you'll meet in AI and machine learning, explained in everyday language.

Phase 5 · lesson 45 min

Object detection: boxes, overlap and suppression

Every object and where it is: boxes in two formats, intersection over union, why one object gets many boxes, non-maximum suppression and its threshold, and grading with average precision.

Never trained an AI? Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account. Teach it to write