{"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/horizonnet-learning-room-layout-with-1d","title":"HorizonNet: Learning Room Layout with 1D Representation and Pano Stretch Data Augmentation","arxiv_id":"1901.03861","date":"2019-01-12","proceeding":"CVPR 2019 6","authors":["Cheng Sun","Chi-Wei Hsiao","Min Sun","Hwann-Tzong Chen"],"abstract":"We present a new approach to the problem of estimating the 3D room layout\nfrom a single panoramic image. We represent room layout as three 1D vectors\nthat encode, at each image column, the boundary positions of floor-wall and\nceiling-wall, and the existence of wall-wall boundary. The proposed network,\nHorizonNet, trained for predicting 1D layout, outperforms previous\nstate-of-the-art approaches. The designed post-processing procedure for\nrecovering 3D room layouts from 1D predictions can automatically infer the room\nshape with low computation cost - it takes less than 20ms for a panorama image\nwhile prior works might need dozens of seconds. We also propose Pano Stretch\nData Augmentation, which can diversify panorama data and be applied to other\npanorama-related learning tasks. Due to the limited data available for\nnon-cuboid layout, we relabel 65 general layout from the current dataset for\nfinetuning. Our approach shows good performance on general layouts by\nqualitative results and cross-validation.","url_abs":"http://arxiv.org/abs/1901.03861v2","url_pdf":"http://arxiv.org/pdf/1901.03861v2.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":"horizonnet-learning-room-layout-with-1d","repo_url":"https://github.com/sunset1995/HorizonNet","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"3d-room-layouts-from-a-single-rgb-panorama","task_name":"3D Room Layouts From A Single RGB Panorama"},{"task_slug":"data-augmentation","task_name":"Data Augmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-room-layouts-from-a-single-rgb-panorama-on","task":"3D Room Layouts From A Single RGB Panorama","dataset":"PanoContext","model":"HorizonNet","rank_in_archive_order":4,"of":7,"metrics":{"3DIoU":"82.17"},"uses_additional_data":false},{"leaderboard":"/sota/3d-room-layouts-from-a-single-rgb-panorama-on-3","task":"3D Room Layouts From A Single RGB Panorama","dataset":"Stanford2D3D Panoramic","model":"HorizonNet","rank_in_archive_order":7,"of":9,"metrics":{"3DIoU":"79.79","Corner Error":"0.71","Pixel Error":"2.39"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1901.03861","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}