top of page
Polygon annotation outlines an object‘s boundary with a series of connected vertices that form a closed shape conforming to the object‘s contour.
In 2D images, polygons fit objects more tightly than bounding boxes and are commonly used to generate ground truth for instance and semantic segmentation.
In 3D point clouds, polygons delineate spatial regions and support tasks such as LiDAR perception for autonomous driving.
Polygons offer a strong balance between precision and cost, making them a good fit for tasks that require tighter boundaries than bounding boxes but do not need full pixel-level accuracy.

Polygon annotation: 2D polygon and 3D polygon

bottom of page


