{"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/vconv-dae-deep-volumetric-shape-learning","title":"VConv-DAE: Deep Volumetric Shape Learning Without Object Labels","arxiv_id":"1604.03755","date":"2016-04-13","proceeding":null,"authors":["Abhishek Sharma","Oliver Grau","Mario Fritz"],"abstract":"With the advent of affordable depth sensors, 3D capture becomes more and more\nubiquitous and already has made its way into commercial products. Yet,\ncapturing the geometry or complete shapes of everyday objects using scanning\ndevices (e.g. Kinect) still comes with several challenges that result in noise\nor even incomplete shapes. Recent success in deep learning has shown how to\nlearn complex shape distributions in a data-driven way from large scale 3D CAD\nModel collections and to utilize them for 3D processing on volumetric\nrepresentations and thereby circumventing problems of topology and\ntessellation. Prior work has shown encouraging results on problems ranging from\nshape completion to recognition. We provide an analysis of such approaches and\ndiscover that training as well as the resulting representation are strongly and\nunnecessarily tied to the notion of object labels. Thus, we propose a full\nconvolutional volumetric auto encoder that learns volumetric representation\nfrom noisy data by estimating the voxel occupancy grids. The proposed method\noutperforms prior work on challenging tasks like denoising and shape\ncompletion. We also show that the obtained deep embedding gives competitive\nperformance when used for classification and promising results for shape\ninterpolation.","url_abs":"http://arxiv.org/abs/1604.03755v3","url_pdf":"http://arxiv.org/pdf/1604.03755v3.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":"vconv-dae-deep-volumetric-shape-learning","repo_url":"https://github.com/diskhkme/VCONV_DAE_TF","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"object","task_name":"Object"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1604.03755","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}