Bounding Box

A bounding box refers to the smallest rectangular or cuboidal box that can completely enclose a given set of points or a complex object. The box is defined by minimum and maximum coordinates along each dimension. In the world of AI and computer vision, a bounding box is often used for object detection, where it's used to define the location and size of the identified objects within an image or a video frame.

How Bounding Boxes work

The creation of a bounding box begins by detecting the outermost points of an object along each dimension. If the object is in two-dimensional space (like an image), it will have four limits: top, bottom, left, and right. For three-dimensional objects, it will have six limits adding front and back to it. The corners of the bounding box are then drawn connecting these outermost points, forming a box that fully encompasses the object.

When employing in computer vision tasks, a bounding box is used in combination with machine learning or deep learning algorithms. The algorithm scans the image or video frame and identifies areas it finds to be objects of interest. These areas are then enclosed in a bounding box. Each box provides crucial information, including the coordinates of the box's location, its dimensions, and the class of object it is enclosing when detected. These bounding boxes form the primary output of object detection models and play a vital role in subsequent tasks such as object tracking, action recognition, and much more.

Lakera LLM Security Playbook
Learn how to protect against the most common LLM vulnerabilities

Download this guide to delve into the most common LLM security risks and ways to mitigate them.

Related terms
untouchable mode.
Get started for free.

Lakera Guard protects your LLM applications from cybersecurity risks with a single line of code. Get started in minutes. Become stronger every day.

Join our Slack Community.

Several people are typing about AI/ML security. 
Come join us and 1000+ others in a chat that’s thoroughly SFW.