{"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/learning-image-representations-tied-to-ego","title":"Learning image representations tied to ego-motion","arxiv_id":"1505.02206","date":"2015-05-08","proceeding":"ICCV 2015 12","authors":["Dinesh Jayaraman","Kristen Grauman"],"abstract":"Understanding how images of objects and scenes behave in response to specific\nego-motions is a crucial aspect of proper visual development, yet existing\nvisual learning methods are conspicuously disconnected from the physical source\nof their images. We propose to exploit proprioceptive motor signals to provide\nunsupervised regularization in convolutional neural networks to learn visual\nrepresentations from egocentric video. Specifically, we enforce that our\nlearned features exhibit equivariance i.e. they respond predictably to\ntransformations associated with distinct ego-motions. With three datasets, we\nshow that our unsupervised feature learning approach significantly outperforms\nprevious approaches on visual recognition and next-best-view prediction tasks.\nIn the most challenging test, we show that features learned from video captured\non an autonomous driving platform improve large-scale scene recognition in\nstatic images from a disjoint domain.","url_abs":"http://arxiv.org/abs/1505.02206v2","url_pdf":"http://arxiv.org/pdf/1505.02206v2.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":"learning-image-representations-tied-to-ego","repo_url":"https://github.com/tu-rbo/concarne","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"autonomous-driving","task_name":"Autonomous Driving"},{"task_slug":"scene-recognition","task_name":"Scene Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1505.02206","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}