Deep Fisher networks for large-scale image classification
As massively parallel computations have become broadly available with modern GPUs, deep architectures trained on very large datasets have risen in popularity. Discriminatively trained convolutional neural networks, in particular, were recently shown to yield state-of-the-art performance in challengi...
| Principais autores: | , , |
|---|---|
| Formato: | Conference item |
| Publicado em: |
Association for Computing Machinery
2013
|
| _version_ | 1826289379365093376 |
|---|---|
| author | Simonyan, K Vedaldi, A Zisserman, A |
| author_facet | Simonyan, K Vedaldi, A Zisserman, A |
| author_sort | Simonyan, K |
| collection | OXFORD |
| description | As massively parallel computations have become broadly available with modern GPUs, deep architectures trained on very large datasets have risen in popularity. Discriminatively trained convolutional neural networks, in particular, were recently shown to yield state-of-the-art performance in challenging image classification benchmarks such as ImageNet. However, elements of these architectures are similar to standard hand-crafted representations used in computer vision. In this paper, we explore the extent of this analogy, proposing a version of the state-of-the-art Fisher vector image encoding that can be stacked in multiple layers. This architecture significantly improves on standard Fisher vectors, and obtains competitive results with deep convolutional networks at a smaller computational learning cost. Our hybrid architecture allows us to assess how the performance of a conventional hand-crafted image classification pipeline changes with increased depth. We also show that convolutional networks and Fisher vector encodings are complementary in the sense that their combination further improves the accuracy. |
| first_indexed | 2024-03-07T02:27:59Z |
| format | Conference item |
| id | oxford-uuid:a647a6c5-96e6-4229-944b-df928a64a282 |
| institution | University of Oxford |
| last_indexed | 2024-03-07T02:27:59Z |
| publishDate | 2013 |
| publisher | Association for Computing Machinery |
| record_format | dspace |
| spelling | oxford-uuid:a647a6c5-96e6-4229-944b-df928a64a2822022-03-27T02:46:10ZDeep Fisher networks for large-scale image classificationConference itemhttp://purl.org/coar/resource_type/c_5794uuid:a647a6c5-96e6-4229-944b-df928a64a282Symplectic Elements at OxfordAssociation for Computing Machinery2013Simonyan, KVedaldi, AZisserman, AAs massively parallel computations have become broadly available with modern GPUs, deep architectures trained on very large datasets have risen in popularity. Discriminatively trained convolutional neural networks, in particular, were recently shown to yield state-of-the-art performance in challenging image classification benchmarks such as ImageNet. However, elements of these architectures are similar to standard hand-crafted representations used in computer vision. In this paper, we explore the extent of this analogy, proposing a version of the state-of-the-art Fisher vector image encoding that can be stacked in multiple layers. This architecture significantly improves on standard Fisher vectors, and obtains competitive results with deep convolutional networks at a smaller computational learning cost. Our hybrid architecture allows us to assess how the performance of a conventional hand-crafted image classification pipeline changes with increased depth. We also show that convolutional networks and Fisher vector encodings are complementary in the sense that their combination further improves the accuracy. |
| spellingShingle | Simonyan, K Vedaldi, A Zisserman, A Deep Fisher networks for large-scale image classification |
| title | Deep Fisher networks for large-scale image classification |
| title_full | Deep Fisher networks for large-scale image classification |
| title_fullStr | Deep Fisher networks for large-scale image classification |
| title_full_unstemmed | Deep Fisher networks for large-scale image classification |
| title_short | Deep Fisher networks for large-scale image classification |
| title_sort | deep fisher networks for large scale image classification |
| work_keys_str_mv | AT simonyank deepfishernetworksforlargescaleimageclassification AT vedaldia deepfishernetworksforlargescaleimageclassification AT zissermana deepfishernetworksforlargescaleimageclassification |