{"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/adversarial-autoencoders-for-generating-3d","title":"Adversarial Autoencoders for Compact Representations of 3D Point Clouds","arxiv_id":"1811.07605","date":"2018-11-19","proceeding":null,"authors":["Maciej Zamorski","Maciej Zięba","Piotr Klukowski","Rafał Nowak","Karol Kurach","Wojciech Stokowiec","Tomasz Trzciński"],"abstract":"Deep generative architectures provide a way to model not only images but also\ncomplex, 3-dimensional objects, such as point clouds. In this work, we present\na novel method to obtain meaningful representations of 3D shapes that can be\nused for challenging tasks including 3D points generation, reconstruction,\ncompression, and clustering. Contrary to existing methods for 3D point cloud\ngeneration that train separate decoupled models for representation learning and\ngeneration, our approach is the first end-to-end solution that allows to\nsimultaneously learn a latent space of representation and generate 3D shape out\nof it. Moreover, our model is capable of learning meaningful compact binary\ndescriptors with adversarial training conducted on a latent space. To achieve\nthis goal, we extend a deep Adversarial Autoencoder model (AAE) to accept 3D\ninput and create 3D output. Thanks to our end-to-end training regime, the\nresulting method called 3D Adversarial Autoencoder (3dAAE) obtains either\nbinary or continuous latent space that covers a much wider portion of training\ndata distribution. Finally, our quantitative evaluation shows that 3dAAE\nprovides state-of-the-art results for 3D points clustering and 3D object\nretrieval.","url_abs":"http://arxiv.org/abs/1811.07605v3","url_pdf":"http://arxiv.org/pdf/1811.07605v3.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":"adversarial-autoencoders-for-generating-3d","repo_url":"https://github.com/MaciejZamorski/3d-AAE","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"adversarial-autoencoders-for-generating-3d","repo_url":"https://github.com/luke9642/master-thesis","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"adversarial-autoencoders-for-generating-3d","repo_url":"https://github.com/texsmv/point_cloud_reconstruction","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"adversarial-autoencoders-for-generating-3d","repo_url":"https://github.com/theamaya/3dlatnav","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"3d-object-retrieval","task_name":"3D Object Retrieval"},{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"generating-3d-point-clouds","task_name":"Generating 3D Point Clouds"},{"task_slug":"point-cloud-generation","task_name":"Point Cloud Generation"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.07605","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}