Deep Metric Learning via Lifted Structured Feature Embedding
Learning the distance metric between pairs of examples is of great importance for learning and visual recognition. With the remarkable success from the state of the art convolutional neural networks, recent works [1, 31] have shown promising results on discriminatively training the networks to learn...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2018
|
| Online Access: | http://hdl.handle.net/1721.1/113397 https://orcid.org/0000-0002-6121-9474 |
| _version_ | 1826209690301759488 |
|---|---|
| author | Song, Hyun Oh Xiang, Yu Savarese, Silvio Jegelka, Stefanie Sabrina |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Song, Hyun Oh Xiang, Yu Savarese, Silvio Jegelka, Stefanie Sabrina |
| author_sort | Song, Hyun Oh |
| collection | MIT |
| description | Learning the distance metric between pairs of examples is of great importance for learning and visual recognition. With the remarkable success from the state of the art convolutional neural networks, recent works [1, 31] have shown promising results on discriminatively training the networks to learn semantic feature embeddings where similar examples are mapped close to each other and dissimilar examples are mapped farther apart. In this paper, we describe an algorithm for taking full advantage of the training batches in the neural network training by lifting the vector of pairwise distances within the batch to the matrix of pairwise distances. This step enables the algorithm to learn the state of the art feature embedding by optimizing a novel structured prediction objective on the lifted problem. Additionally, we collected Stanford Online Products dataset: 120k images of 23k classes of online products for metric learning. Our experiments on the CUB-200-2011 [37], CARS196 [19], and Stanford Online Products datasets demonstrate significant improvement over existing deep feature embedding methods on all experimented embedding sizes with the GoogLeNet [33] network. The source code and the dataset are available at: https://github.com/rksltnl/ Deep-Metric-Learning-CVPR16. |
| first_indexed | 2024-09-23T14:27:17Z |
| format | Article |
| id | mit-1721.1/113397 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T14:27:17Z |
| publishDate | 2018 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1133972022-10-01T21:25:31Z Deep Metric Learning via Lifted Structured Feature Embedding Song, Hyun Oh Xiang, Yu Savarese, Silvio Jegelka, Stefanie Sabrina Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Jegelka, Stefanie Sabrina Learning the distance metric between pairs of examples is of great importance for learning and visual recognition. With the remarkable success from the state of the art convolutional neural networks, recent works [1, 31] have shown promising results on discriminatively training the networks to learn semantic feature embeddings where similar examples are mapped close to each other and dissimilar examples are mapped farther apart. In this paper, we describe an algorithm for taking full advantage of the training batches in the neural network training by lifting the vector of pairwise distances within the batch to the matrix of pairwise distances. This step enables the algorithm to learn the state of the art feature embedding by optimizing a novel structured prediction objective on the lifted problem. Additionally, we collected Stanford Online Products dataset: 120k images of 23k classes of online products for metric learning. Our experiments on the CUB-200-2011 [37], CARS196 [19], and Stanford Online Products datasets demonstrate significant improvement over existing deep feature embedding methods on all experimented embedding sizes with the GoogLeNet [33] network. The source code and the dataset are available at: https://github.com/rksltnl/ Deep-Metric-Learning-CVPR16. United States. Office of Naval Research (grant #N00014-13- 1-0761) Stanford University. SAIL-Toyota Center for AI Research (grant #122282) 2018-02-02T18:20:36Z 2018-02-02T18:20:36Z 2016-06 Article http://purl.org/eprint/type/ConferencePaper 978-1-4673-8851-1 1063-6919 http://hdl.handle.net/1721.1/113397 Song, Hyun Oh, Yu Xiang, Stefanie Jegelka, and Silvio Savarese. “Deep Metric Learning via Lifted Structured Feature Embedding.” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (June 2016). https://orcid.org/0000-0002-6121-9474 en_US http://dx.doi.org/10.1109/CVPR.2016.434 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) arXiv |
| spellingShingle | Song, Hyun Oh Xiang, Yu Savarese, Silvio Jegelka, Stefanie Sabrina Deep Metric Learning via Lifted Structured Feature Embedding |
| title | Deep Metric Learning via Lifted Structured Feature Embedding |
| title_full | Deep Metric Learning via Lifted Structured Feature Embedding |
| title_fullStr | Deep Metric Learning via Lifted Structured Feature Embedding |
| title_full_unstemmed | Deep Metric Learning via Lifted Structured Feature Embedding |
| title_short | Deep Metric Learning via Lifted Structured Feature Embedding |
| title_sort | deep metric learning via lifted structured feature embedding |
| url | http://hdl.handle.net/1721.1/113397 https://orcid.org/0000-0002-6121-9474 |
| work_keys_str_mv | AT songhyunoh deepmetriclearningvialiftedstructuredfeatureembedding AT xiangyu deepmetriclearningvialiftedstructuredfeatureembedding AT savaresesilvio deepmetriclearningvialiftedstructuredfeatureembedding AT jegelkastefaniesabrina deepmetriclearningvialiftedstructuredfeatureembedding |