Panoptic segmentation mask is a mask that integrates semantic and instance segmentation, marking pixels by semantic categories (e.g., road, pedestrian) and distinguishing different target instances under the same category, enabling full-scene pixel-level annotation of images for panoptic scene understanding.
