{"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/eliminating-the-blind-spot-adapting-3d-object-1","title":"Eliminating the Blind Spot: Adapting 3D Object Detection and Monocular Depth Estimation to 360° Panoramic Imagery","arxiv_id":"1808.06253","date":"2018-08-19","proceeding":"ECCV 2018","authors":["Grégoire Payen de La Garanderie","Amir Atapour Abarghouei","Toby P. Breckon"],"abstract":"Recent automotive vision work has focused almost exclusively on processing\nforward-facing cameras. However, future autonomous vehicles will not be viable\nwithout a more comprehensive surround sensing, akin to a human driver, as can\nbe provided by 360{\\deg} panoramic cameras. We present an approach to adapt\ncontemporary deep network architectures developed on conventional rectilinear\nimagery to work on equirectangular 360{\\deg} panoramic imagery. To address the\nlack of annotated panoramic automotive datasets availability, we adapt a\ncontemporary automotive dataset, via style and projection transformations, to\nfacilitate the cross-domain retraining of contemporary algorithms for panoramic\nimagery. Following this approach we retrain and adapt existing architectures to\nrecover scene depth and 3D pose of vehicles from monocular panoramic imagery\nwithout any panoramic training labels or calibration parameters. Our approach\nis evaluated qualitatively on crowd-sourced panoramic images and quantitatively\nusing an automotive environment simulator to provide the first benchmark for\nsuch techniques within panoramic imagery.","url_abs":"http://arxiv.org/abs/1808.06253v1","url_pdf":"http://arxiv.org/pdf/1808.06253v1.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":"eliminating-the-blind-spot-adapting-3d-object-1","repo_url":"https://github.com/gdlg/panoramic-depth-estimation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"3d-object-detection","task_name":"3D Object Detection"},{"task_slug":"autonomous-vehicles","task_name":"Autonomous Vehicles"},{"task_slug":"depth-estimation","task_name":"Depth Estimation"},{"task_slug":"monocular-depth-estimation","task_name":"Monocular Depth Estimation"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1808.06253","atlas_url":"https://app.syntology.ai/?focus=1808.06253","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}