{"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-interpretable-non-rigid-structure-from","title":"Deep Interpretable Non-Rigid Structure from Motion","arxiv_id":"1902.10840","date":"2019-02-28","proceeding":null,"authors":["Chen Kong","Simon Lucey"],"abstract":"All current non-rigid structure from motion (NRSfM) algorithms are limited\nwith respect to: (i) the number of images, and (ii) the type of shape\nvariability they can handle. This has hampered the practical utility of NRSfM\nfor many applications within vision. In this paper we propose a novel deep\nneural network to recover camera poses and 3D points solely from an ensemble of\n2D image coordinates. The proposed neural network is mathematically\ninterpretable as a multi-layer block sparse dictionary learning problem, and\ncan handle problems of unprecedented scale and shape complexity. Extensive\nexperiments demonstrate the impressive performance of our approach where we\nexhibit superior precision and robustness against all available\nstate-of-the-art works. The considerable model capacity of our approach affords\nremarkable generalization to unseen data. We propose a quality measure (based\non the network weights) which circumvents the need for 3D ground-truth to\nascertain the confidence we have in the reconstruction. Once the network's\nweights are estimated (for a non-rigid object) we show how our approach can\neffectively recover 3D shape from a single image -- outperforming comparable\nmethods that rely on direct 3D supervision.","url_abs":"http://arxiv.org/abs/1902.10840v1","url_pdf":"http://arxiv.org/pdf/1902.10840v1.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-interpretable-non-rigid-structure-from","repo_url":"https://github.com/kongchen1992/deep-nrsfm","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"dictionary-learning","task_name":"Dictionary Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}