A bounding box is a rectangular boundary drawn around an object or region of interest, most commonly in computer vision and image processing. Its four sides indicate the approximate location and size of the detected object within an image or video frame. For example, an object-detection model may place a bounding box around each vehicle it identifies in a road scene, allowing software to determine where those vehicles are and track them over time.
Bounding boxes are widely used for training and evaluating object-detection models. They provide simpler object localization than pixel-level segmentation, making them useful for applications such as autonomous driving, surveillance, robotics, and automated visual inspection.
