Global and Local Features Based Classification for Bleed-Through Removal

Xiangyu Hu; Hui Lin; Shutao Li; Bin Sun

首页> 外文期刊>Sensing and imaging >Global and Local Features Based Classification for Bleed-Through Removal

【24h】

Global and Local Features Based Classification for Bleed-Through Removal

机译：基于全局和局部特征的流血去除分类

获取原文

获取原文并翻译 | 示例

掌桥外文数据库（机构版） >>

开具论文收录证明 >>

文献代查 >>

页面导航

摘要
著录项
相似文献
相关主题

摘要

The text on one side of historical documents often seeps through andrnappears on the other side, so the bleed-through is a common problem in historicalrndocument images. It makes the document images hard to read and the text difficultrnto recognize. To improve the image quality and readability, the bleed-through has tornbe removed. This paper proposes a global and local features extraction based bleedthroughrnremoval method. The Gaussian mixture model is used to get the globalrnfeatures of the images. Local features are extracted by the patch around each pixel.rnThen, the extreme learning machine classifier is utilized to classify the scannedrnimages into the foreground text and the bleed-through component. Experimentalrnresults on real document image datasets show that the proposed method outperformsrnthe state-of-the-art bleed-through removal methods and preserves the text strokesrnwell.

机译：历史文档一侧的文本通常会在另一侧渗透并出现，因此渗漏是历史文档图像中的常见问题。它使文档图像难以阅读，文本难以识别。为了提高图像质量和可读性，已撕掉直通孔。本文提出了一种基于全局和局部特征提取的渗漏去除方法。高斯混合模型用于获得图像的全局特征。通过每个像素周围的补丁提取局部特征。然后，利用极限学习机分类器将扫描图像分类为前景文本和渗漏成分。对真实文档图像数据集的实验结果表明，所提出的方法优于最先进的渗漏消除方法，并能很好地保留文本笔触。

著录项

来源
《Sensing and imaging》 |2016年第1期|9.1-9.14|共14页
作者
Xiangyu Hu; Hui Lin; Shutao Li; Bin Sun;
展开▼
作者单位

College of Electrical and Information Engineering, Hunan University, Changsha 410082, China;

College of Electrical and Information Engineering, Hunan University, Changsha 410082, China;

College of Electrical and Information Engineering, Hunan University, Changsha 410082, China;

College of Electrical and Information Engineering, Hunan University, Changsha 410082, China;

展开▼
收录信息
原文格式 PDF
正文语种 eng
中图分类
关键词
Classification; Bleed-through; Extreme learning machine; Features extraction; Scanned document image;

机译：分类;流血;极限学习机;特征提取;扫描文件图像;

相似文献

外文文献
中文文献
专利

1. Global-feature classification can be acquired more rapidly than local-feature classification in both humans and pigeons [J] . Goto K, Wills AJ, Lea SEG Animal Cognition . 2004,第2期

机译：在人类和鸽子中，全局特征分类都可以比局部特征分类更快地获取
2. Non-Local Sparse Image Inpainting for Document Bleed-Through Removal [J] . Muhammad Hanif, Anna Tonazzini, Pasquale Savino, Journal of Imaging . 2018,第5期

机译：非局部稀疏图像修复，可去除文档的流血
3. Semi-supervised process fault classification based on convolutional ladder network with local and global feature fusion [J] . Shipeng Li, Jiaxiang Luo, Yueming Hu Computers & Chemical Engineering . 2020,第Sepa2期

机译：基于卷积梯网络与本地和全局特征融合的半监控过程故障分类
4. CNN Classification Based on Global and Local Features [C] . Yufeng Zheng, Jun Huang, Tianwen Chen, SPIE Defense + Commercial Sensing Conference . 2019

机译：基于全局和本地特征的CNN分类
5. Exploring image and video by classification and clustering on global and local visual features. [D] . Lu, Le. 2007

机译：通过对全局和局部视觉特征进行分类和聚类来探索图像和视频。
6. Many Local Pattern Texture Features: Which Is Better for Image-Based Multilabel Human Protein Subcellular Localization Classification? [O] . Fan Yang, Ying-Ying Xu, Hong-Bin Shen -1

机译：许多局部图案纹理特征：哪种更好的基于图像的多标签人类蛋白质亚细胞定位分类？
7. A Novel Technique of Feature Extraction Based on Local and Global Similarity Measure for Protein Classification [O] . Neha Bharill, Aruna Tiwari 2015

机译：一种基于局部和全球相似性蛋白质分类的特征提取技术

Global and Local Features Based Classification for Bleed-Through Removal

摘要

著录项

相似文献

相关主题

期刊订阅