{"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/real-time-deep-video-deinterlacing","title":"Real-time Deep Video Deinterlacing","arxiv_id":"1708.00187","date":"2017-08-01","proceeding":null,"authors":["Haichao Zhu","Xueting Liu","Xiangyu Mao","Tien-Tsin Wong"],"abstract":"Interlacing is a widely used technique, for television broadcast and video\nrecording, to double the perceived frame rate without increasing the bandwidth.\nBut it presents annoying visual artifacts, such as flickering and silhouette\n\"serration,\" during the playback. Existing state-of-the-art deinterlacing\nmethods either ignore the temporal information to provide real-time performance\nbut lower visual quality, or estimate the motion for better deinterlacing but\nwith a trade-off of higher computational cost. In this paper, we present the\nfirst and novel deep convolutional neural networks (DCNNs) based method to\ndeinterlace with high visual quality and real-time performance. Unlike existing\nmodels for super-resolution problems which relies on the translation-invariant\nassumption, our proposed DCNN model utilizes the temporal information from both\nthe odd and even half frames to reconstruct only the missing scanlines, and\nretains the given odd and even scanlines for producing the full deinterlaced\nframes. By further introducing a layer-sharable architecture, our system can\nachieve real-time performance on a single GPU. Experiments shows that our\nmethod outperforms all existing methods, in terms of reconstruction accuracy\nand computational performance.","url_abs":"http://arxiv.org/abs/1708.00187v1","url_pdf":"http://arxiv.org/pdf/1708.00187v1.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":"real-time-deep-video-deinterlacing","repo_url":"https://github.com/gordinirojo/deinterlace","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"real-time-deep-video-deinterlacing","repo_url":"https://github.com/lszhuhaichao/Deep-Video-Deinterlacing","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":null,"task_name":"GPU"},{"task_slug":"super-resolution","task_name":"Super-Resolution"},{"task_slug":"translation","task_name":"Translation"},{"task_slug":"video-deinterlacing","task_name":"Video Deinterlacing"}],"methods":[{"method_slug":"dcnn","method_name":"DCNN"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/video-deinterlacing-on-msu-deinterlacer","task":"Video Deinterlacing","dataset":"MSU Deinterlacer Benchmark","model":"Real-time Deep Video Deinterlacing","rank_in_archive_order":11,"of":31,"metrics":{"FPS on CPU":"0.3","PSNR":"38.374","SSIM":"0.957","Subjective":"0.543","VMAF":"93.28"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}