Bounding Box

A bounding box is typically defined using coordinates that specify its position, width, and height within an image. During object detection, AI models place these boxes around detected objects and often assign each box a class label and confidence score. Bounding boxes can also be added manually during data annotation to create labeled datasets for training computer vision models.

Bounding boxes help AI systems determine both what an object is and where it appears. They provide a relatively simple and efficient way to train and evaluate object detection models without requiring precise pixel-level annotations.