{"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/self-view-grounding-given-a-narrated-360","title":"Self-view Grounding Given a Narrated 360° Video","arxiv_id":"1711.08664","date":"2017-11-23","proceeding":null,"authors":["Shih-Han Chou","Yi-Chun Chen","Kuo-Hao Zeng","Hou-Ning Hu","Jianlong Fu","Min Sun"],"abstract":"Narrated 360{\\deg} videos are typically provided in many touring scenarios to\nmimic real-world experience. However, previous work has shown that smart\nassistance (i.e., providing visual guidance) can significantly help users to\nfollow the Normal Field of View (NFoV) corresponding to the narrative. In this\nproject, we aim at automatically grounding the NFoVs of a 360{\\deg} video given\nsubtitles of the narrative (referred to as \"NFoV-grounding\"). We propose a\nnovel Visual Grounding Model (VGM) to implicitly and efficiently predict the\nNFoVs given the video content and subtitles. Specifically, at each frame, we\nefficiently encode the panorama into feature map of candidate NFoVs using a\nConvolutional Neural Network (CNN) and the subtitles to the same hidden space\nusing an RNN with Gated Recurrent Units (GRU). Then, we apply soft-attention on\ncandidate NFoVs to trigger sentence decoder aiming to minimize the reconstruct\nloss between the generated and given sentence. Finally, we obtain the NFoV as\nthe candidate NFoV with the maximum attention without any human supervision. To\ntrain VGM more robustly, we also generate a reverse sentence conditioning on\none minus the soft-attention such that the attention focuses on candidate NFoVs\nless relevant to the given sentence. The negative log reconstruction loss of\nthe reverse sentence (referred to as \"irrelevant loss\") is jointly minimized to\nencourage the reverse sentence to be different from the given sentence. To\nevaluate our method, we collect the first narrated 360{\\deg} videos dataset and\nachieve state-of-the-art NFoV-grounding performance.","url_abs":"http://arxiv.org/abs/1711.08664v1","url_pdf":"http://arxiv.org/pdf/1711.08664v1.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":"self-view-grounding-given-a-narrated-360","repo_url":"https://github.com/ShihHanChou/360grounding","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"visual-grounding","task_name":"Visual Grounding"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.08664","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}