{"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/perspective-transformer-nets-learning-single","title":"Perspective Transformer Nets: Learning Single-View 3D Object Reconstruction without 3D Supervision","arxiv_id":"1612.00814","date":"2016-12-01","proceeding":"NeurIPS 2016 12","authors":["Xinchen Yan","Jimei Yang","Ersin Yumer","Yijie Guo","Honglak Lee"],"abstract":"Understanding the 3D world is a fundamental problem in computer vision.\nHowever, learning a good representation of 3D objects is still an open problem\ndue to the high dimensionality of the data and many factors of variation\ninvolved. In this work, we investigate the task of single-view 3D object\nreconstruction from a learning agent's perspective. We formulate the learning\nprocess as an interaction between 3D and 2D representations and propose an\nencoder-decoder network with a novel projection loss defined by the perspective\ntransformation. More importantly, the projection loss enables the unsupervised\nlearning using 2D observation without explicit 3D supervision. We demonstrate\nthe ability of the model in generating 3D volume from a single 2D image with\nthree sets of experiments: (1) learning from single-class objects; (2) learning\nfrom multi-class objects and (3) testing on novel object classes. Results show\nsuperior performance and better generalization ability for 3D object\nreconstruction when the projection loss is involved.","url_abs":"http://arxiv.org/abs/1612.00814v3","url_pdf":"http://arxiv.org/pdf/1612.00814v3.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":"perspective-transformer-nets-learning-single","repo_url":"https://github.com/xcyan/nips16_PTN","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"torch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"perspective-transformer-nets-learning-single","repo_url":"https://github.com/tensorflow/models/tree/master/research/ptn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"3d-object-reconstruction","task_name":"3D Object Reconstruction"},{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-reconstruction","task_name":"Object Reconstruction"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1612.00814","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}