{"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/learning-to-cluster-for-proposal-free","title":"Learning to Cluster for Proposal-Free Instance Segmentation","arxiv_id":"1803.06459","date":"2018-03-17","proceeding":null,"authors":["Yen-Chang Hsu","Zheng Xu","Zsolt Kira","Jiawei Huang"],"abstract":"This work proposed a novel learning objective to train a deep neural network\nto perform end-to-end image pixel clustering. We applied the approach to\ninstance segmentation, which is at the intersection of image semantic\nsegmentation and object detection. We utilize the most fundamental property of\ninstance labeling -- the pairwise relationship between pixels -- as the\nsupervision to formulate the learning objective, then apply it to train a fully\nconvolutional network (FCN) for learning to perform pixel-wise clustering. The\nresulting clusters can be used as the instance labeling directly. To support\nlabeling of an unlimited number of instance, we further formulate ideas from\ngraph coloring theory into the proposed learning objective. The evaluation on\nthe Cityscapes dataset demonstrates strong performance and therefore proof of\nthe concept. Moreover, our approach won the second place in the lane detection\ncompetition of 2017 CVPR Autonomous Driving Challenge, and was the top\nperformer without using external data.","url_abs":"http://arxiv.org/abs/1803.06459v1","url_pdf":"http://arxiv.org/pdf/1803.06459v1.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":"learning-to-cluster-for-proposal-free","repo_url":"https://github.com/GT-RIPL/L2C","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"autonomous-driving","task_name":"Autonomous Driving"},{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"instance-segmentation","task_name":"Instance Segmentation"},{"task_slug":"lane-detection","task_name":"Lane Detection"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/lane-detection-on-tusimple","task":"Lane Detection","dataset":"TuSimple","model":"Pairwise pixel supervision + FCN","rank_in_archive_order":17,"of":43,"metrics":{"Accuracy":"96.50%","F1 score":"94.31"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1803.06459","atlas_url":"https://app.syntology.ai/?focus=1803.06459","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}