Mask annotation is an annotation operation that generates masks through pixel-level marking (e.g., category ID, instance ID) in image segmentation tasks, accurately outlining target contours and regions, providing training data for semantic, instance, and panoptic segmentation models.
