Bounding Box Annotation

Drawing rectangles around objects in images — the bread-and-butter of computer-vision training.

Definition

Bounding-box annotation draws axis-aligned rectangles around target objects (cars, people, traffic signs) so the AI learns to detect them. Specialty variants: 3D cuboids for LiDAR, polygon segmentation for pixel-precise edges.

Examples

  • Boxing every pedestrian in a driving video
  • Boxing tumors in a medical scan

Related terms

Looking for actual gigs that use Bounding Box Annotation? Browse all live listings or filter by $100+/hr.