检索结果-内蒙古大学图书馆

An enhanced binarization framework for degraded historical document images

EURASIP JOURNAL ON image AND VIDEO PROCESSING 2021年第1期2021卷 1-24页

作者： Xiong, Wei Zhou, Lei Yue, Ling Li, Lirong Wang, Song Hubei Univ Technol Sch Elect & Elect Engn Wuhan 430068 Hubei Peoples R China Univ South Carolina Dept Comp Sci & Engn Columbia SC 29208 USA

Binarization plays an important role in document analysis and recognition (DAR) systems. In this paper, we present our winning algorithm in ICFHR 2018 competition on handwritten document image binarization (H-DIBCO 2018), which is based on background estimation and energy minimization. First, we adopt mathematical morphological operations to estimate and compensate the document background. It uses a disk-shaped structuring element, whose radius is computed by the minimum entropy-based stroke width transform (SWT). Second, we perform Laplacian energy-based segmentation on the compensated document images. Finally, we implement post-processing to preserve text stroke connectivity and eliminate isolated noise. Experimental results indicate that the proposed method outperforms other state-of-the-art techniques on several public available benchmark datasets.

关键词： document image binarization document image segmentation Background estimation and compensation Laplacian energy minimization Minimum entropy-based stroke width transform (SWT) Markov random fields (MRFs) Graph cut

来源：评论

学校读者我要写书评

暂无评论

Texture sparseness for pixel classification of business document images

引用

INTERNATIONAL JOURNAL ON document ANALYSIS AND RECOGNITION 2014年第3期17卷 257-273页

作者： Cote, Melissa Albu, Alexandra Branzan Univ Victoria STN CSC Victoria BC V8W 3P6 Canada

Contemporary business documents contain diverse, multi-layered mixtures of textual, graphical, and pictorial elements. Existing methods for document segmentation and classification do not handle well the complexity and variety of contents, geometric layout, and elemental shapes. This paper proposes a novel document image classification approach that distributes individual pixels into four fundamental classes (text, image, graphics, and background) through support vector machines. This approach uses a novel low-dimensional feature descriptor based on textural properties. The proposed feature vector is constructed by considering the sparseness of the document image responses to a filter bank on a multi-resolution and contextual basis. Qualitative and quantitative evaluations on business document images show the benefits of adopting a contextual and multi-resolution approach. The proposed approach achieves excellent results;it is able to handle varied contents and complex document layouts, without imposing any constraint or making assumptions about the shape and spatial arrangement of document elements.

关键词： Business documents document image segmentation Pixel classification Sparseness Support vector machines Texture

来源：评论

学校读者我要写书评

暂无评论

Context-based multiscale classification of document images using wavelet coefficient distributions

引用

IEEE TRANSACTIONS ON image PROCESSING 2000年第9期9卷 1604-1616页

作者： Li, J Gray, RM Xerox Corp Palo Alto Res Ctr Palo Alto CA 94304 USA Stanford Univ Dept Elect Engn Informat Syst Lab Stanford CA 94305 USA

In this paper, an algorithm is developed for segmenting document images into four classes: background, photograph, text, acid graph. Features used for classification are based on the distribution patterns of wavelet coefficients in high frequency bands. Two important attributes of the algorithm are its multiscale nature-it classifies an image at different resolutions adaptively, enabling accurate classification at class boundaries as well as fast classification overall-and its use of accumulated context information for improving classification accuracy.

关键词： context-dependent classification document image segmentation goodness of match multiscale classification text and photograph segmentation wavelet transform

来源：评论

学校读者我要写书评

暂无评论

Blind Decomposition of Multispectral document images Using Orthogonal Nonnegative Matrix Factorization

引用

IEEE TRANSACTIONS ON image PROCESSING 2021年 30卷 5997-6012页

作者： Rahiche, Abderrahmane Cheriet, Mohamed Univ Quebec Synchromedia Lab Ecole Technol Super ETS Montreal PQ H3C 1K3 Canada

This paper addresses the challenge of Multispectral (MS) document image segmentation, which is an essential step for subsequent document image analysis. Most previous studies have focused only on binary (text/non-text) separation. They also rely on handcrafted features and techniques dedicated to conventional images that do not take advantage of MS images' spectral richness. In this work, we reformulate this task as a source separation problem, whereby we target the blind decomposition of entire MS document images via a new orthogonal nonnegative matrix factorization (ONMF). On the one hand, we incorporate orthogonality constraint as a Riemannian optimization on the Stiefel manifold. On the other hand, based on which factor we impose the orthogonality constraint, i.e., either on the endmember matrix, abundance matrix, or both, we propose three ONMF models to investigate this issue and determine which model is more suitable for this study. Minimizing the three models subject to nonnegativity and orthogonality constraints simultaneously is very challenging. Therefore, we extend the alternating direction method of multipliers scheme to solve them. We evaluated our models on synthetic Hyperspectral (HS) images and real-world MS document images. The experimental results confirm the effectiveness of the proposed models and demonstrate their generalization power compared to state-of-the-art techniques.

关键词： Manifolds Optimization Data models Matrix decomposition Task analysis image segmentation image analysis Multispectral images orthogonal nonnegative matrix factorization Stiefel manifold document image segmentation Blind source separation

来源：评论

学校读者我要写书评

暂无评论

Loss Function for Training Models of segmentation of document images

引用

PROGRAMMING AND COMPUTER SOFTWARE 2023年第7期49卷 574-589页

作者： Perminov, A. I. Turdakov, D. Yu. Belyaeva, O. V. Russian Acad Sci Ivannikov Inst Syst Programming Moscow 109004 Russia Moscow MV Lomonosov State Univ Moscow 119991 Russia

This work is devoted to improving the quality of segmentation of images of various scientific papers and legal acts by neural network models by training them using modified loss functions that take into account special features of images of the appropriate subject domain. The analysis of existing loss functions is carried out, and new functions are proposed that work both with the coordinates of bounding boxes and use information about the pixels of the input image. To assess the quality, a neural network segmentation model with modified loss functions is trained, and a theoretical assessment is carried out using a simulation experiment showing the convergence rate and segmentation error. As a result of the study, rapidly converging loss functions are created that improve the quality of document image segmentation using additional information about the input data.

关键词： document image segmentation loss functions modification of loss function

来源：评论

学校读者我要写书评

暂无评论

document image binarisation based on low-rank modelling

引用

ELECTRONICS LETTERS 2015年第22期51卷 1784-1785页

作者： Zheng, Xin Chen, Yung-Sheng Xiamen Univ Dept Elect Sci Fujian Prov Key Lab Plasma & Magnet Resonance Xiamen Peoples R China Yuan Ze Univ Dept Elect Engn Taoyuan Taiwan

segmentation from document images with a complex background is presented. The background component is estimated by low-rank modelling. With the estimated background component, the original image is further transformed into a compensated image which is easily segmented with Otsu's thresholding algorithm to obtain a good binarisation result. Experiments on the Bickley diary and the DIBCO 2013 datasets validate the feasibility of the proposed method.

关键词： document image processing image segmentation Bickley diary DIBCO 2013 datasets Otsu thresholding algorithm background component compensated image complex background document image binarisation document image segmentation low rank modelling

来源：评论

学校读者我要写书评

暂无评论

Simple and effective table detection system from document images

引用

INTERNATIONAL JOURNAL ON document ANALYSIS AND RECOGNITION 2006年第2-3期8卷 172-182页

作者： Mandal, S. Chowdhury, S. P. Das, A. K. Chanda, Bhabatosh Bengal Engn Coll DU CST Dept Howrah 711103 W Bengal India Indian Stat Unit ECS Unit Kolkata 700035 W Bengal India

The requirement of detection and identification of tables from document images is crucial to any document image analysis and digital library system. In this paper we report a very simple but extremely powerful approach to detect tables present in document pages. The algorithm relies on the observation that the tables have distinct columns which implies that gaps between the fields are substantially larger than the gaps between the words in text lines. This deceptively simple observation has led to the design of a simple but powerful table detection system with low computation cost. Moreover, mathematical foundation of the approach is also established including formation of a regular expression for ease of implementation.

关键词： table detection document image segmentation digital document library

来源：评论

学校读者我要写书评

暂无评论

Neuro-fuzzy segmentation of document images

Neuro-fuzzy segmentation of document images

引用

5th IASTED International Conference on Visualization, Imaging, and image Processing

作者： Górecki, P Castiello, C Caponetti, L Univ Bari Dipartimento Informat I-70126 Bari Italy

ISBN: (纸本)0889865280

The task of document image segmentation is to represent a digital image in a more interpretable form, recognising regions containing text, background and graphics. This paper presents a peculiar strategy for document image segmentation, where a neuro-fuzzy approach is involved. Firstly, image is segmented into text, graphics or background during a pixel level classification step. Successively, an analysis performed over the obtained regions is devoted to refine the initial segmentation results. A knowledge discovery process is applied to automatically derive from sample data the fuzzy rule bases, responsible of the inference scheme presiding over the classification of image pixels and regions. The proposed method proves to be accurate and robust to page skew and noise.

关键词： document analysis neuro-fuzzy classification document image segmentation feature extraction

来源：评论

学校读者我要写书评

暂无评论

Palmira: A Deep Deformable Network for Instance segmentation of Dense and Uneven Layouts in Handwritten Manuscripts 16th

Palmira: A Deep Deformable Network for Instance Segmentation...

引用

16th IAPR International Conference on document Analysis and Recognition (ICDAR)

作者： Sharan, S. P. Aitha, Sowmya Kumar, Amandeep Trivedi, Abhishek Augustine, Aaron Sarvadevabhatla, Ravi Kiran Int Inst Informat Technol Ctr Visual Informat Technol Hyderabad India

ISBN: (纸本)9783030863319

Handwritten documents are often characterized by dense and uneven layout. Despite advances, standard deep network based approaches for semantic layout segmentation are not robust to complex deformations seen across semantic regions. This phenomenon is especially pronounced for the low-resource Indic palm-leaf manuscript domain. To address the issue, we first introduce Indiscapes2, a new large-scale diverse dataset of Indic manuscripts with semantic layout annotations. Indiscapes2 contains documents from four different historical collections and is 150% larger than its predecessor, Indiscapes. We also propose a novel deep network PALMIRA for robust, deformation-aware instance segmentation of regions in handwritten manuscripts. We also report Hausdorff distance and its variants as a boundary-aware performance measure. Our experiments demonstrate that PALMIRA provides robust layouts, outperforms strong baseline approaches and ablative variants. We also include qualitative results on Arabic, South-East Asian and Hebrew historical manuscripts to showcase the generalization capability of PALMIRA.

关键词： Instance segmentation Deformable convolutional network Historical document analysis document image segmentation Dataset

来源：评论

学校读者我要写书评

暂无评论

HPSegNet: A Method for Handwritten and Printed Text Separation in document images

HPSegNet: A Method for Handwritten and Printed Text Separati...

引用

18th International Conference on document Analysis and Recognition (ICDAR)

作者： Chao, Yu Liu, Changsong Peng, Liangrui Wang, Yanwei Tsinghua Univ Dept Elect Engn Beijing Peoples R China Beijing Natl Res Ctr Informat Sci & Technol BNRis Beijing Peoples R China

ISBN: (纸本)9783031706417;9783031706424

The separation of handwritten and printed text in document images is an important task in the optical character recognition (OCR) research field. It is still a challenging problem to separate overlapped handwritten and printed text lines in images of complex documents including examination papers, legal documents, etc. In this paper, hand-written and printed text separation is formulated as a pixel-level document image segmentation task. Firstly, a modified Transformer-based model is designed for pixel-level document image segmentation. Secondly, a residual feature bypass is incorporated into the model to further exploit high-resolution features. Finally, a loss function combining focal loss and dice loss is designed to tackle the problem of imbalanced distributions of different classes. Experimental results on both a public English document image dataset and a self-built Chinese document image dataset have demonstrated the effectiveness of the proposed method.

关键词： Handwritten Text Printed Text document image segmentation Deep Learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：