Implementação oficial de Pytorch do detector de texto artesanal | Papel | Modelo pré -terenciado | Suplementar
Youngmin Baek, Bado Lee, Dongyoon Han, Sangdoo Yun, Hwalsuk Lee.
Clova AI Research, Naver Corp.
Implementação de Pytorch para detector de texto artesanato que detecta efetivamente a área de texto, explorando cada região de cada caractere e afinidade entre os caracteres. A caixa delimitadora de textos é obtida simplesmente encontrando retângulos mínimos delimitadores no mapa binário após a região do caractere de limiares e as pontuações de afinidade.

13 de junho de 2019 : Atualização inicial 20 de julho de 2019 : Adicionado pós-processamento para o resultado do polígono 28 de setembro de 2019 : Adicionado o modelo treinado no IC15 e o link refinador
pip install -r requirements.txt
O código para treinamento não está incluído neste repositório e não podemos liberar o código de treinamento completo por motivo de IP.
| Nome do modelo | Conjuntos de dados usados | Idiomas | Propósito | Modelo Link |
|---|---|---|---|---|
| Em geral | SynthText, IC13, IC17 | Eng + mlt | Para uso geral | Clique |
| IC15 | SynthText, IC15 | Eng | Apenas para IC15 | Clique |
| LinkRefiner | CTW1500 | - | Usado com o modelo geral | Clique |
python test.py --trained_model=[weightfile] --test_folder=[folder path to test images]
A imagem de resultado e os mapas Socre serão salvos para ./result por padrão.
--trained_model : modelo pré-terenciado--text_threshold : limite de confiança do texto--low_text : pontuação de texto baixo de texto--link_threshold : Link Confidence Limfil--cuda : Use Cuda para inferência (padrão: true)--canvas_size : tamanho máximo de imagem para inferência--mag_ratio : taxa de ampliação da imagem--poly : Ativar resultado do tipo polígono--show_time : mostre tempo de processamento--test_folder : Path de pasta para inserir imagens--refine : use Link Refiner para o conjunto de dados de nível Sentense--refiner_model : modelo de refinador pré-terenciado @inproceedings{baek2019character,
title={Character Region Awareness for Text Detection},
author={Baek, Youngmin and Lee, Bado and Han, Dongyoon and Yun, Sangdoo and Lee, Hwalsuk},
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
pages={9365--9374},
year={2019}
}
Copyright (c) 2019-present NAVER Corp.
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.