Semantic label sharing for learning with many categories
In an object recognition scenario with tens of thousands of categories, even a small number of labels per category leads to a very large number of total labels required. We propose a simple method of label sharing between semantically similar categories. We leverage the WordNet hierarchy to define s...
Main Authors: | , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | en_US |
Published: |
Association for Computing Machinery
2011
|
Online Access: | http://hdl.handle.net/1721.1/64732 https://orcid.org/0000-0003-4915-0256 |
_version_ | 1811089313004257280 |
---|---|
author | Fergus, Rob Bernal, Hector Weiss, Yair Torralba, Antonio |
author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Fergus, Rob Bernal, Hector Weiss, Yair Torralba, Antonio |
author_sort | Fergus, Rob |
collection | MIT |
description | In an object recognition scenario with tens of thousands of categories, even a small number of labels per category leads to a very large number of total labels required. We propose a simple method of label sharing between semantically similar categories. We leverage the WordNet hierarchy to define semantic distance between any two categories and use this semantic distance to share labels. Our approach can be used with any classifier. Experimental results on a range of datasets, up to 80 million images and 75,000 categories in size, show that despite the simplicity of the approach, it leads to significant improvements in performance. |
first_indexed | 2024-09-23T14:17:13Z |
format | Article |
id | mit-1721.1/64732 |
institution | Massachusetts Institute of Technology |
language | en_US |
last_indexed | 2024-09-23T14:17:13Z |
publishDate | 2011 |
publisher | Association for Computing Machinery |
record_format | dspace |
spelling | mit-1721.1/647322022-09-28T19:44:53Z Semantic label sharing for learning with many categories Fergus, Rob Bernal, Hector Weiss, Yair Torralba, Antonio Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Torralba, Antonio Torralba, Antonio Bernal, Hector In an object recognition scenario with tens of thousands of categories, even a small number of labels per category leads to a very large number of total labels required. We propose a simple method of label sharing between semantically similar categories. We leverage the WordNet hierarchy to define semantic distance between any two categories and use this semantic distance to share labels. Our approach can be used with any classifier. Experimental results on a range of datasets, up to 80 million images and 75,000 categories in size, show that despite the simplicity of the approach, it leads to significant improvements in performance. 2011-06-30T19:41:28Z 2011-06-30T19:41:28Z 2010-09 Article http://purl.org/eprint/type/ConferencePaper 978-3-642-15548-2 3-642-15548-0 http://hdl.handle.net/1721.1/64732 Fergus, Rob et al. “Semantic Label Sharing for Learning with Many Categories.” in Proceedings of the 11th European Conference on Computer Vision: Part I. Heraklion, Crete, Sept. 5-11, Greece: Springer-Verlag, 2010. 762-775. (Lecture Notes in Computer Science, Vol. 6311) https://orcid.org/0000-0003-4915-0256 en_US http://portal.acm.org/citation.cfm?id=1886121 Proceedings of the 11th European conference on Computer vision: Part I, ECCV'10 Creative Commons Attribution-Noncommercial-Share Alike 3.0 http://creativecommons.org/licenses/by-nc-sa/3.0/ application/pdf Association for Computing Machinery MIT web domain |
spellingShingle | Fergus, Rob Bernal, Hector Weiss, Yair Torralba, Antonio Semantic label sharing for learning with many categories |
title | Semantic label sharing for learning with many categories |
title_full | Semantic label sharing for learning with many categories |
title_fullStr | Semantic label sharing for learning with many categories |
title_full_unstemmed | Semantic label sharing for learning with many categories |
title_short | Semantic label sharing for learning with many categories |
title_sort | semantic label sharing for learning with many categories |
url | http://hdl.handle.net/1721.1/64732 https://orcid.org/0000-0003-4915-0256 |
work_keys_str_mv | AT fergusrob semanticlabelsharingforlearningwithmanycategories AT bernalhector semanticlabelsharingforlearningwithmanycategories AT weissyair semanticlabelsharingforlearningwithmanycategories AT torralbaantonio semanticlabelsharingforlearningwithmanycategories |