论文阅读 [TPAMI-2022] DE-GAN: A Conditional Generative Adversarial Network for Document Enhancement
论文阅读 [TPAMI-2022] DE-GAN: A Conditional Generative Adversarial Network for Document Enhancement
论文搜索(studyai.com)
搜索论文: DE-GAN: A Conditional Generative Adversarial Network for Document Enhancement
搜索论文: http://www.studyai.com/search/whole-site/?q=DE-GAN:+A+Conditional+Generative+Adversarial+Network+for+Document+Enhancement
关键字(Keywords)
Generative adversarial networks; Text analysis; Machine learning; Image restoration; Document analysis; document enhancement; degraded document binarization; watermark removal; deep learning; generative adversarial networks
机器视觉
生成对抗; 图像/视频修复
摘要(Abstract)
Documents often exhibit various forms of degradation, which make it hard to be read and substantially deteriorate the performance of an OCR system.
文档通常会表现出各种形式的降级，这使其难以读取，并严重影响OCR系统的性能。.
In this paper, we propose an effective end-to-end framework named document enhancement generative adversarial networks (DE-GAN) that uses the conditional GANs (cGANs) to restore severely degraded document images.
在本文中，我们提出了一个有效的端到端框架，名为文档增强生成对抗网络（DE-GAN），它使用条件GAN（CGAN）来恢复严重退化的文档图像。.
To the best of our knowledge, this practice has not been studied within the context of generative adversarial deep networks.
据我们所知，这种做法尚未在生成性对抗性深层网络的背景下进行研究。.
We demonstrate that, in different tasks (document clean up, binarization, deblurring and watermark removal), DE-GAN can produce an enhanced version of the degraded document with a high quality.
我们证明，在不同的任务（文档清理、二值化、去模糊和水印删除）中，DE-GAN可以生成高质量的降级文档的增强版本。.
In addition, our approach provides consistent improvements compared to state-of-the-art methods over the widely used DIBCO 2013, DIBCO 2017, and H-DIBCO 2018 datasets, proving its ability to restore a degraded document image to its ideal condition.
此外，与广泛使用的DIBCO 2013、DIBCO 2017和H-DIBCO 2018数据集相比，我们的方法提供了一致的改进，证明了它能够将降级的文档图像恢复到理想状态。.
The obtained results on a wide variety of degradation reveal the flexibility of the proposed model to be exploited in other document enhancement problems…
在各种退化情况下获得的结果显示了所提出模型的灵活性，可用于其他文档增强问题。。.
作者(Authors)
[‘Mohamed Ali Souibgui’, ‘Yousri Kessentini’]