{"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/matryoshka-networks-predicting-3d-geometry","title":"Matryoshka Networks: Predicting 3D Geometry via Nested Shape Layers","arxiv_id":"1804.10975","date":"2018-04-29","proceeding":"CVPR 2018 6","authors":["Stephan R. Richter","Stefan Roth"],"abstract":"In this paper, we develop novel, efficient 2D encodings for 3D geometry,\nwhich enable reconstructing full 3D shapes from a single image at high\nresolution. The key idea is to pose 3D shape reconstruction as a 2D prediction\nproblem. To that end, we first develop a simple baseline network that predicts\nentire voxel tubes at each pixel of a reference view. By leveraging well-proven\narchitectures for 2D pixel-prediction tasks, we attain state-of-the-art\nresults, clearly outperforming purely voxel-based approaches. We scale this\nbaseline to higher resolutions by proposing a memory-efficient shape encoding,\nwhich recursively decomposes a 3D shape into nested shape layers, similar to\nthe pieces of a Matryoshka doll. This allows reconstructing highly detailed\nshapes with complex topology, as demonstrated in extensive experiments; we\nclearly outperform previous octree-based approaches despite having a much\nsimpler architecture using standard network components. Our Matryoshka networks\nfurther enable reconstructing shapes from IDs or shape similarity, as well as\nshape sampling.","url_abs":"http://arxiv.org/abs/1804.10975v1","url_pdf":"http://arxiv.org/pdf/1804.10975v1.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":"matryoshka-networks-predicting-3d-geometry","repo_url":"https://bitbucket.org/visinf/projects-2018-matryoshka","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"matryoshka-networks-predicting-3d-geometry","repo_url":"https://github.com/JeremyFisher/deep_level_sets","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"matryoshka-networks-predicting-3d-geometry","repo_url":"https://github.com/JeremyFisher/few_shot_3dr","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"3d-object-reconstruction","task_name":"3D Object Reconstruction"},{"task_slug":"3d-shape-reconstruction","task_name":"3D Shape Reconstruction"},{"task_slug":"3d-geometry","task_name":"3D geometry"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-object-reconstruction-on-data3dr2n2","task":"3D Object Reconstruction","dataset":"Data3D−R2N2","model":"Matryoshka Networks","rank_in_archive_order":7,"of":15,"metrics":{"3DIoU":"0.640"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.10975","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}