{"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/deep-360-pilot-learning-a-deep-agent-for","title":"Deep 360 Pilot: Learning a Deep Agent for Piloting through 360° Sports Video","arxiv_id":"1705.01759","date":"2017-05-04","proceeding":"CVPR 2017","authors":["Hou-Ning Hu","Yen-Chen Lin","Ming-Yu Liu","Hsien-Tzu Cheng","Yung-Ju Chang","Min Sun"],"abstract":"Watching a 360{\\deg} sports video requires a viewer to continuously select a\nviewing angle, either through a sequence of mouse clicks or head movements. To\nrelieve the viewer from this \"360 piloting\" task, we propose \"deep 360 pilot\"\n-- a deep learning-based agent for piloting through 360{\\deg} sports videos\nautomatically. At each frame, the agent observes a panoramic image and has the\nknowledge of previously selected viewing angles. The task of the agent is to\nshift the current viewing angle (i.e. action) to the next preferred one (i.e.,\ngoal). We propose to directly learn an online policy of the agent from data. We\nuse the policy gradient technique to jointly train our pipeline: by minimizing\n(1) a regression loss measuring the distance between the selected and ground\ntruth viewing angles, (2) a smoothness loss encouraging smooth transition in\nviewing angle, and (3) maximizing an expected reward of focusing on a\nforeground object. To evaluate our method, we build a new 360-Sports video\ndataset consisting of five sports domains. We train domain-specific agents and\nachieve the best performance on viewing angle selection accuracy and transition\nsmoothness compared to [51] and other baselines.","url_abs":"http://arxiv.org/abs/1705.01759v1","url_pdf":"http://arxiv.org/pdf/1705.01759v1.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":"deep-360-pilot-learning-a-deep-agent-for","repo_url":"https://github.com/eborboihuc/Deep360Pilot-CVPR17","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.01759","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}