Multimodal AI
Bounding box
A rectangle marking the location and approximate extent of an object in an image.
Example
An object detector draws a box around a vehicle.
Why people use it
It marks roughly where an object appears so software can count or follow it.
What you'll hear
“Put a rectangle around each vehicle.”
What this means for you
Use segmentation when precise pixel boundaries are needed.
Can you control it?
Developer-only
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Can one object receive two boxes by mistake?
- Yes. A detector can produce overlapping duplicate guesses that need to be combined or removed.
- Does the whole rectangle belong to the object?
- No. The box usually includes some background around the object's shape.
- Can overlapping boxes refer to different objects?
- Yes. Objects can overlap in a picture while still having separate boxes.