{"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/pushing-the-boundaries-of-view-extrapolation","title":"Pushing the Boundaries of View Extrapolation with Multiplane Images","arxiv_id":"1905.00413","date":"2019-05-01","proceeding":"CVPR 2019 6","authors":["Pratul P. Srinivasan","Richard Tucker","Jonathan T. Barron","Ravi Ramamoorthi","Ren Ng","Noah Snavely"],"abstract":"We explore the problem of view synthesis from a narrow baseline pair of\nimages, and focus on generating high-quality view extrapolations with plausible\ndisocclusions. Our method builds upon prior work in predicting a multiplane\nimage (MPI), which represents scene content as a set of RGB$\\alpha$ planes\nwithin a reference view frustum and renders novel views by projecting this\ncontent into the target viewpoints. We present a theoretical analysis showing\nhow the range of views that can be rendered from an MPI increases linearly with\nthe MPI disparity sampling frequency, as well as a novel MPI prediction\nprocedure that theoretically enables view extrapolations of up to $4\\times$ the\nlateral viewpoint movement allowed by prior work. Our method ameliorates two\nspecific issues that limit the range of views renderable by prior methods: 1)\nWe expand the range of novel views that can be rendered without depth\ndiscretization artifacts by using a 3D convolutional network architecture along\nwith a randomized-resolution training procedure to allow our model to predict\nMPIs with increased disparity sampling frequency. 2) We reduce the repeated\ntexture artifacts seen in disocclusions by enforcing a constraint that the\nappearance of hidden content at any depth must be drawn from visible content at\nor behind that depth. Please see our results video at:\nhttps://www.youtube.com/watch?v=aJqAaMNL2m4.","url_abs":"http://arxiv.org/abs/1905.00413v1","url_pdf":"http://arxiv.org/pdf/1905.00413v1.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":"pushing-the-boundaries-of-view-extrapolation","repo_url":"https://github.com/google-research/google-research/tree/master/mpi_extrapolation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1905.00413","atlas_url":"https://app.syntology.ai/?focus=1905.00413","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}