# Learning Instance Occlusion for Panoptic Segmentation Source: [Learning Instance Occlusion for Panoptic Segmentation]() Collection index: [llms.txt]() Data updated: 2026-01-04T03:09:42.371Z Authors: Justin Lazarow, Kwonjoon Lee, Kunyu Shi, Zhuowen Tu Published: 2019-06-13T19:04:58.000Z ## Abstract Panoptic segmentation requires segments of both "things" (countable object instances) and "stuff" (uncountable and amorphous regions) within a single output. A common approach involves the fusion of instance segmentation (for "things") and semantic segmentation (for "stuff") into a non-overlapping placement of segments, and resolves overlaps. However, instance ordering with detection confidence do not correlate well with natural occlusion relationship. To resolve this issue, we propose a branch that is tasked with modeling how two instance masks should overlap one another as a binary relation. Our method, named OCFusion, is lightweight but particularly effective in the instance fusion process. OCFusion is trained with the ground truth relation derived automatically from the existing dataset annotations. We obtain state-of-the-art results on COCO and show competitive results on the Cityscapes panoptic segmentation benchmark. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Panoptic Segmentation on MS-COCO (val)](): Panoptic Segmentation; MS-COCO (val); PQ: 46.3. - [Panoptic Segmentation on COCO (test-dev)](): Panoptic Segmentation; COCO (test-dev); PQ: 46.7. - [Panoptic Segmentation on COCO (val)](): Panoptic Segmentation; COCO (val); PQ: 41.2. - [Panoptic Segmentation on Cityscapes (val)](): Panoptic Segmentation; Cityscapes (val); PQ: 60.2. ## Original sources and implementation [Original paper]() [PDF]()