{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/single-network-panoptic-segmentation-for","title":"Single Network Panoptic Segmentation for Street Scene Understanding","arxiv_id":"1902.02678","date":"2019-02-07","proceeding":null,"authors":["Daan de Geus","Panagiotis Meletis","Gijs Dubbelman"],"abstract":"In this work, we propose a single deep neural network for panoptic\nsegmentation, for which the goal is to provide each individual pixel of an\ninput image with a class label, as in semantic segmentation, as well as a\nunique identifier for specific objects in an image, following instance\nsegmentation. Our network makes joint semantic and instance segmentation\npredictions and combines these to form an output in the panoptic format. This\nhas two main benefits: firstly, the entire panoptic prediction is made in one\npass, reducing the required computation time and resources; secondly, by\nlearning the tasks jointly, information is shared between the two tasks,\nthereby improving performance. Our network is evaluated on two street scene\ndatasets: Cityscapes and Mapillary Vistas. By leveraging information exchange\nand improving the merging heuristics, we increase the performance of the single\nnetwork, and achieve a score of 23.9 on the Panoptic Quality (PQ) metric on\nMapillary Vistas validation, with an input resolution of 640 x 900 pixels. On\nCityscapes validation, our method achieves a PQ score of 45.9 with an input\nresolution of 512 x 1024 pixels. Moreover, our method decreases the prediction\ntime by a factor of 2 with respect to separate networks.","url_abs":"http://arxiv.org/abs/1902.02678v1","url_pdf":"http://arxiv.org/pdf/1902.02678v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"single-network-panoptic-segmentation-for","repo_url":"https://github.com/DdeGeus/single-network-panoptic-segmentation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"instance-segmentation","task_name":"Instance Segmentation"},{"task_slug":"panoptic-segmentation","task_name":"Panoptic Segmentation"},{"task_slug":"scene-understanding","task_name":"Scene Understanding"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1902.02678","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}