{"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/horizon-lines-in-the-wild","title":"Horizon Lines in the Wild","arxiv_id":"1604.02129","date":"2016-04-07","proceeding":null,"authors":["Scott Workman","Menghua Zhai","Nathan Jacobs"],"abstract":"The horizon line is an important contextual attribute for a wide variety of\nimage understanding tasks. As such, many methods have been proposed to estimate\nits location from a single image. These methods typically require the image to\ncontain specific cues, such as vanishing points, coplanar circles, and regular\ntextures, thus limiting their real-world applicability. We introduce a large,\nrealistic evaluation dataset, Horizon Lines in the Wild (HLW), containing\nnatural images with labeled horizon lines. Using this dataset, we investigate\nthe application of convolutional neural networks for directly estimating the\nhorizon line, without requiring any explicit geometric constraints or other\nspecial cues. An extensive evaluation shows that using our CNNs, either in\nisolation or in conjunction with a previous geometric approach, we achieve\nstate-of-the-art results on the challenging HLW dataset and two existing\nbenchmark datasets.","url_abs":"http://arxiv.org/abs/1604.02129v2","url_pdf":"http://arxiv.org/pdf/1604.02129v2.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":"horizon-lines-in-the-wild","repo_url":"https://github.com/scottworkman/deephorizon","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"caffe2","reach":null}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"horizon-line-estimation","task_name":"Horizon Line Estimation"}],"methods":[],"datasets_introduced":[{"slug":"hlw","name":"HLW","full_name":"Horizon Lines in the Wild"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/horizon-line-estimation-on-eurasian-cities","task":"Horizon Line Estimation","dataset":"Eurasian Cities Dataset","model":"GoogleNet (Huber Loss, horizon line projection)","rank_in_archive_order":4,"of":4,"metrics":{"AUC (horizon error)":"83.6"},"uses_additional_data":false},{"leaderboard":"/sota/horizon-line-estimation-on-horizon-lines-in","task":"Horizon Line Estimation","dataset":"Horizon Lines in the Wild","model":"GoogleNet (Huber Loss, horizon line projection)","rank_in_archive_order":2,"of":5,"metrics":{"AUC (horizon error)":"71.16"},"uses_additional_data":false},{"leaderboard":"/sota/horizon-line-estimation-on-york-urban-dataset","task":"Horizon Line Estimation","dataset":"York Urban Dataset","model":"GoogleNet (Huber Loss, horizon line projection)","rank_in_archive_order":4,"of":4,"metrics":{"AUC (horizon error)":"86.41"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1604.02129","atlas_url":"https://app.syntology.ai/?focus=1604.02129","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}