{"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/a-skeleton-bridged-deep-learning-approach-for","title":"A Skeleton-bridged Deep Learning Approach for Generating Meshes of Complex Topologies from Single RGB Images","arxiv_id":"1903.04704","date":"2019-03-12","proceeding":"CVPR 2019 6","authors":["Jiapeng Tang","Xiaoguang Han","Junyi Pan","Kui Jia","Xin Tong"],"abstract":"This paper focuses on the challenging task of learning 3D object surface\nreconstructions from single RGB images. Existing methods achieve varying\ndegrees of success by using different geometric representations. However, they\nall have their own drawbacks, and cannot well reconstruct those surfaces of\ncomplex topologies. To this end, we propose in this paper a skeleton-bridged,\nstage-wise learning approach to address the challenge. Our use of skeleton is\ndue to its nice property of topology preservation, while being of lower\ncomplexity to learn. To learn skeleton from an input image, we design a deep\narchitecture whose decoder is based on a novel design of parallel streams\nrespectively for synthesis of curve- and surface-like skeleton points. We use\ndifferent shape representations of point cloud, volume, and mesh in our\nstage-wise learning, in order to take their respective advantages. We also\npropose multi-stage use of the input image to correct prediction errors that\nare possibly accumulated in each stage. We conduct intensive experiments to\ninvestigate the efficacy of our proposed approach. Qualitative and quantitative\nresults on representative object categories of both simple and complex\ntopologies demonstrate the superiority of our approach over existing ones. We\nwill make our ShapeNet-Skeleton dataset publicly available.","url_abs":"http://arxiv.org/abs/1903.04704v2","url_pdf":"http://arxiv.org/pdf/1903.04704v2.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":"a-skeleton-bridged-deep-learning-approach-for","repo_url":"https://github.com/tangjiapeng/SkeletonBridgeRecon","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"}],"methods":[{"method_slug":"affine-coupling","method_name":"Affine Coupling"},{"method_slug":"normalizing-flows","method_name":"Normalizing Flows"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1903.04704","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}