Synthetic data and artificial neural networks for natural scene text recognition
In this work we present a framework for the recognition of natural scene text. Our framework does not require any human-labelled data, and performs word recognition on the whole image holistically, departing from the character based recognition systems of the past. The deep neural network models at...
| Main Authors: | , , , |
|---|---|
| Format: | Conference item |
| Published: |
Neural Information Processing Systems
2014
|
| _version_ | 1826298373842403328 |
|---|---|
| author | Jaderberg, M Simonyan, K Vedaldi, A Zisserman, A |
| author_facet | Jaderberg, M Simonyan, K Vedaldi, A Zisserman, A |
| author_sort | Jaderberg, M |
| collection | OXFORD |
| description | In this work we present a framework for the recognition of natural scene text. Our framework does not require any human-labelled data, and performs word recognition on the whole image holistically, departing from the character based recognition systems of the past. The deep neural network models at the centre of this framework are trained solely on data produced by a synthetic text generation engine – synthetic data that is highly realistic and sufficient to replace real data, giving us infinite amounts of training data. This excess of data exposes new possibilities for word recognition models, and here we consider three models, each one “reading” words in a different way: via 90k-way dictionary encoding, character sequence encoding, and bag-of-N-grams encoding. In the scenarios of language based and completely unconstrained text recognition we greatly improve upon state-of-the-art performance on standard datasets, using our fast, simple machinery and requiring zero data-acquisition costs. |
| first_indexed | 2024-03-07T04:45:51Z |
| format | Conference item |
| id | oxford-uuid:d33ca010-8b8f-48d5-ab4b-cc449540598d |
| institution | University of Oxford |
| last_indexed | 2024-03-07T04:45:51Z |
| publishDate | 2014 |
| publisher | Neural Information Processing Systems |
| record_format | dspace |
| spelling | oxford-uuid:d33ca010-8b8f-48d5-ab4b-cc449540598d2022-03-27T08:09:49ZSynthetic data and artificial neural networks for natural scene text recognitionConference itemhttp://purl.org/coar/resource_type/c_5794uuid:d33ca010-8b8f-48d5-ab4b-cc449540598dSymplectic Elements at OxfordNeural Information Processing Systems2014Jaderberg, MSimonyan, KVedaldi, AZisserman, AIn this work we present a framework for the recognition of natural scene text. Our framework does not require any human-labelled data, and performs word recognition on the whole image holistically, departing from the character based recognition systems of the past. The deep neural network models at the centre of this framework are trained solely on data produced by a synthetic text generation engine – synthetic data that is highly realistic and sufficient to replace real data, giving us infinite amounts of training data. This excess of data exposes new possibilities for word recognition models, and here we consider three models, each one “reading” words in a different way: via 90k-way dictionary encoding, character sequence encoding, and bag-of-N-grams encoding. In the scenarios of language based and completely unconstrained text recognition we greatly improve upon state-of-the-art performance on standard datasets, using our fast, simple machinery and requiring zero data-acquisition costs. |
| spellingShingle | Jaderberg, M Simonyan, K Vedaldi, A Zisserman, A Synthetic data and artificial neural networks for natural scene text recognition |
| title | Synthetic data and artificial neural networks for natural scene text recognition |
| title_full | Synthetic data and artificial neural networks for natural scene text recognition |
| title_fullStr | Synthetic data and artificial neural networks for natural scene text recognition |
| title_full_unstemmed | Synthetic data and artificial neural networks for natural scene text recognition |
| title_short | Synthetic data and artificial neural networks for natural scene text recognition |
| title_sort | synthetic data and artificial neural networks for natural scene text recognition |
| work_keys_str_mv | AT jaderbergm syntheticdataandartificialneuralnetworksfornaturalscenetextrecognition AT simonyank syntheticdataandartificialneuralnetworksfornaturalscenetextrecognition AT vedaldia syntheticdataandartificialneuralnetworksfornaturalscenetextrecognition AT zissermana syntheticdataandartificialneuralnetworksfornaturalscenetextrecognition |