{"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/shape-completion-using-3d-encoder-predictor","title":"Shape Completion using 3D-Encoder-Predictor CNNs and Shape Synthesis","arxiv_id":"1612.00101","date":"2016-12-01","proceeding":"CVPR 2017 7","authors":["Angela Dai","Charles Ruizhongtai Qi","Matthias Nießner"],"abstract":"We introduce a data-driven approach to complete partial 3D shapes through a\ncombination of volumetric deep neural networks and 3D shape synthesis. From a\npartially-scanned input shape, our method first infers a low-resolution -- but\ncomplete -- output. To this end, we introduce a 3D-Encoder-Predictor Network\n(3D-EPN) which is composed of 3D convolutional layers. The network is trained\nto predict and fill in missing data, and operates on an implicit surface\nrepresentation that encodes both known and unknown space. This allows us to\npredict global structure in unknown areas at high accuracy. We then correlate\nthese intermediary results with 3D geometry from a shape database at test time.\nIn a final pass, we propose a patch-based 3D shape synthesis method that\nimposes the 3D geometry from these retrieved shapes as constraints on the\ncoarsely-completed mesh. This synthesis process enables us to reconstruct\nfine-scale detail and generate high-resolution output while respecting the\nglobal mesh structure obtained by the 3D-EPN. Although our 3D-EPN outperforms\nstate-of-the-art completion method, the main contribution in our work lies in\nthe combination of a data-driven shape predictor and analytic 3D shape\nsynthesis. In our results, we show extensive evaluations on a newly-introduced\nshape completion benchmark for both real-world and synthetic data.","url_abs":"http://arxiv.org/abs/1612.00101v2","url_pdf":"http://arxiv.org/pdf/1612.00101v2.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":"shape-completion-using-3d-encoder-predictor","repo_url":"https://github.com/angeladai/cnncomplete","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"torch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"shape-completion-using-3d-encoder-predictor","repo_url":"https://github.com/zzxmllq/cnn_com","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"3d-shape-generation","task_name":"3D Shape Generation"},{"task_slug":"3d-geometry","task_name":"3D geometry"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1612.00101","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}