Image Segmentation
A computer vision task that assigns a class label to every pixel in an image, dividing it into meaningful regions.
Types
Semantic segmentation: Labels every pixel with a class (sky, road, car) but doesn't distinguish between instances. Instance segmentation: Identifies individual objects (car 1, car 2). Panoptic segmentation: Combines both -- labels everything and distinguishes instances.
Key Models
U-Net (medical imaging standard), Mask R-CNN (instance segmentation), SAM (Segment Anything Model from Meta -- segments any object with zero-shot prompting), and DeepLab (semantic segmentation).
Applications
Autonomous driving (road/object parsing), medical imaging (tumor boundaries), satellite imagery analysis, video editing (background removal), and augmented reality.