{"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/labelfusion-a-pipeline-for-generating-ground","title":"LabelFusion: A Pipeline for Generating Ground Truth Labels for Real RGBD Data of Cluttered Scenes","arxiv_id":"1707.04796","date":"2017-07-15","proceeding":null,"authors":["Pat Marion","Peter R. Florence","Lucas Manuelli","Russ Tedrake"],"abstract":"Deep neural network (DNN) architectures have been shown to outperform\ntraditional pipelines for object segmentation and pose estimation using RGBD\ndata, but the performance of these DNN pipelines is directly tied to how\nrepresentative the training data is of the true data. Hence a key requirement\nfor employing these methods in practice is to have a large set of labeled data\nfor your specific robotic manipulation task, a requirement that is not\ngenerally satisfied by existing datasets. In this paper we develop a pipeline\nto rapidly generate high quality RGBD data with pixelwise labels and object\nposes. We use an RGBD camera to collect video of a scene from multiple\nviewpoints and leverage existing reconstruction techniques to produce a 3D\ndense reconstruction. We label the 3D reconstruction using a human assisted\nICP-fitting of object meshes. By reprojecting the results of labeling the 3D\nscene we can produce labels for each RGBD image of the scene. This pipeline\nenabled us to collect over 1,000,000 labeled object instances in just a few\ndays. We use this dataset to answer questions related to how much training data\nis required, and of what quality the data must be, to achieve high performance\nfrom a DNN architecture.","url_abs":"http://arxiv.org/abs/1707.04796v3","url_pdf":"http://arxiv.org/pdf/1707.04796v3.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":"labelfusion-a-pipeline-for-generating-ground","repo_url":"https://github.com/RobotLocomotion/LabelFusion","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"3d-reconstruction","task_name":"3D Reconstruction"},{"task_slug":"object","task_name":"Object"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1707.04796","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}