top of page
Image segmentation is a pixel-level data labeling task in computer vision.
A model assigns every pixel in an image a class or region label, dividing the image into meaningful regions and producing precise object or region boundaries.
It is a core building block for applications such as autonomous driving, robotics, medical imaging, remote sensing, and AR.
Common forms include semantic segmentation, instance segmentation, and panoptic segmentation, which differ in how they treat object instances and background classes.
Example of image segmentation
bottom of page