Comparative Study of Text Line Segmentation Algorithms on Low Quality Documents

Text line segmentation is a necessary preliminary step before most text transcription algorithms are applied. The leading deep learning networks used in this context (ARU-Net, dhSegment, and Doc-UFCN) are based on the U-Net architecture. They are efficient, but fall under the same concept, requiring a post-processing step to perform instance (e.g., text line) segmentation. In the present work, we test the advantages of Mask-RCNN, which is designed to perform instance segmentation directly. This work is the first to directly compare Mask-RCNN- and U-Net-based networks on text segmentation of historical documents, showing the superiority of the former over the latter. Three studies were conducted, one comparing these networks on different historical databases, another comparing Mask-RCNN with Doc-UFCN on a private historical database, and a third comparing the handwritten text recognition (HTR) performance of the tested networks. The results showed that Mask-RCNN outperformed ARU-Net, dhSegment, and Doc-UFCN using relevant line segmentation metrics, that performance evaluation should not focus on the raw masks generated by the networks, that a light mask processing is an efficient and simple solution to improve evaluation, and that Mask-RCNN leads to better HTR performance.

A novel grouping approach to segment text lines from handwritten documents is presented. In this text line segmentation algorithm, for each text line, a text string that connects the center points of the characters in this text line is built. The text lines are then segmented using the resulting text strings. Since the characters of the same text line are situated close together and aligned on a smooth curve, 2D tensor voting is used to reduce the conflicts when building these text strings. First, the text lines are represented by separate connected components. The center points of these connected components are then encoded by second order tensors. Finally, a voting process is applied to extract the curve saliency values and normal vectors, which are used to remove outliers and build the text strings. The experimental results obtained from the test dataset of the ICDAR 2009 Handwriting Segmentation Contest show that the proposed method generates high detection rate and recognition accuracy.

Segmentation is a very critical stage in the character recognition process as the performance of any character recognition system depends heavily on the accuracy of segmentation. Although segmentation is a well-researched area, segmentation of handwritten text is still difficult owing to several factors like skewed and overlapping lines, the presence of touching, broken and degraded characters, and variations in writing styles. Therefore, researchers in this area are working continuously to develop new techniques for the efficient segmentation and recognition of characters. In the character recognition process, segmentation can be implemented at the line, word, and character level. Text line segmentation is the first step in the text/character recognition process. The line segmentation methods used in the character recognition of handwritten documents are presented in this paper. The various levels of segmentation which include line, word, and character segmentation are discussed with a focus on line segmentation.KeywordsSegmentationOCRPattern recognitionLine segmentationWord segmentationCharacter segmentationTouching lines segmentationOverlapping lines segmentation

Text line segmentation is an important preprocessing work for Dongba document analysis. It is the basis of analysis and collection work such as character segmentation, Dongba pictograph extraction, Dongba pictograph translation and annotation segmentation, and sample collection. Combining with the structural characteristics of Dongba hieroglyphic documents, we give a text line segmentation algorithm which is suitable for Dongba hieroglyphic documents. The algorithm combines K-means clustering on the basis of projection segmentation to realize the automatic determination of the height range of a single text line. At the same time, it also realizes the automatic segmentation of too large or too small text blocks. This will lay the foundation for the automatic collection of Dongba pictographs and the completion of further research on Dongba documents.

Despite recent advances in the field of supervised deep learning for text line segmentation, unsupervised deep learning solutions are beginning to gain popularity. In this paper, we present an unsupervised deep learning method that embeds document image patches to a compact Euclidean space where distances correspond to a coarse text line pattern similarity. Once this space has been produced, text line segmentation can be easily implemented using standard techniques with the embedded feature vectors. To train the model, we extract random pairs of document image patches with the assumption that neighbour patches contain a similar coarse trend of text lines, whereas if one of them is rotated, they contain different coarse trends of text lines. Doing well on this task requires the model to learn to recognize the text lines and their salient parts. The benefit of our approach is zero manual labelling effort. We evaluate the method qualitatively and quantitatively on several variants of text line segmentation datasets to demonstrate its effectivity.

There are more than 1000 languages and 14 scripts used by 112 million people in India. All of these scripts divide the document in three parts: Text block, Image block, and Table block. In 21st century, there is a need, obvious reasons, to convert these old printed documents in digital form. Converting them manually is a huge and difficult task. Further it is prone to human errors. Another automated technique is to use Optical character recognition (OCR) system to convert the entire printed document image into editable document. In this paper, an effort has been made to develop OCR technique which converts the printed document into editable document. Firstly a scanned document is preprocessed for noise and skew correction. It is then followed by text-non text classification. Then text line detection has to be performed in the text area. There is no method available which can detect the text line if the image contains the multicolumn text area. In this paper the main contribution is to detect the blocks and detect the text lines in these detected blocks. The technique which can extract the text lines in image document is presented here. After extraction of text lines, word segmentation, character segmentation, and template matching can be performed.

Due to the presence of high volume of noise in text documents, it becomes very difficult to achieve high accuracy while performing text line segmentation. Hence, approaches which are dependent on the performance of the line segmentation stage, also suffers. Also due to the variability of noise patterns, the text contents get distorted and sometimes it results in broken strokes of a character. Thus, character level recognition becomes a challenging task for noisy documents. Our proposed method is aimed to address the aforesaid challenges. Since, we are considering the input image to be of noisy kind, a convolutional neural network (CNN) architecture is introduced to initially identify whether the input image contains noise or not. We consider this as a two-class problem and try to identify the respective classes, i.e. noisy or clean. Then, we apply the proposed piecewise projection profile feature and an adaptive region growing based two-stage algorithmic approach which initially identifies the text line upper and lower boundaries in a noisy text document image and then try to regroup the broken strokes of a character to enhance the character recognition accuracy. The proposed method has been tested on a large size dataset containing noisy documents and also different quality measures are computed to establish the effectiveness of the proposed method. From the measures, it can be noted that the text line segmentation accuracy is comparable to other similar state-of-the-art results.

Text line segmentation in unconstrained handwritten documents remains a challenge because handwritten text lines are multi-skewed and not obviously separated. This paper presents a new approach based on the variational Bayes (VB) framework for text line segmentation. Viewing the document image as a mixture density model, with each text line approximated by a Gaussian component, the VB method can automatically determine the number of components. We extend the VB method such that it can both eliminate and split components and control the orientation of text line lines. Experiments on Chinese handwritten documents demonstrated the effectiveness of the approach.

Optical Character Recognition (OCR) is a major computer vision task by which characters of image are detected and recognized by comparing to training set images. Process of detecting character is one of the perplexing tasks in computer vision. This is because of input image often not correctly aligned or because of noise. This paper presents a complete Optical Character Recognition (OCR) system which is worked for English character mostly for Calibri font. This system first corrects skew of image if input image is not correctly aligned followed by noise reduction from input image. This process is passed through line and character segmentation that are passed into the recognition module and recognize characters. By experimenting with a set of 50 images, average achievement is 92%, 98% is for Calibri font. Moreover, the developed technique is computationally efficient and requires less time than other Optical character recognition system.

Text line detection is crucial for any application associated with Automatic Text Recognition or Keyword Spotting. Modern algorithms perform good on well-established datasets since they either comprise clean data or simple/homogeneous page layouts. We have collected and annotated 2036 archival document images from different locations and time periods. The dataset contains varying page layouts and degradations that challenge text line segmentation methods. Well established text line segmentation evaluation schemes such as the Detection Rate or Recognition Accuracy demand for binarized data that is annotated on a pixel level. Producing ground truth by these means is laborious and not needed to determine a method's quality. In this paper we propose a new evaluation scheme that is based on baselines. The proposed scheme has no need for binarization and it can handle skewed as well as rotated text lines. The ICDAR 2017 Competition on Baseline Detection and the ICDAR 2017 Competition on Layout Analysis for Challenging Medieval Manuscripts used this evaluation scheme. Finally, we present results achieved by a recently published text line detection algorithm.

The purpose of this survey is to provide a comprehensive overview of recent advancements in text line segmentation and baseline detection techniques within the analysis of historical document images. Text line extraction is an essential step in the historical documents image analysis pipeline, as its results significantly impact the accuracy of subsequent processes, such as handwritten text recognition (HTR). Through a multi-stage procedure, we carefully selected 49 peer-reviewed studies published since 2019. Based on careful analysis of these studies, we summarize the information of the existing datasets, describe and categorize different methods, and summarize evaluation protocols. In addition, we compare the results of various methods on benchmark datasets. Finally, we highlight the gaps and suggest directions for future research. We believe that this comprehensive survey will be of great assistance to researchers working in the field of historical document image analysis, as it offers critical insights into the latest advancements and developments, providing a foundation for future research.

Text line segmentation (TLS) continues to be an unsolved problem in the field of document analysis, and this is especially true for lines of handwritten characters. In recent years, however, this challenging problem has seen a surge of interest. In this paper we introduce our work on a new robust framework for solving the TLS problem via a trial-and-error method within a divide-and-conquer paradigm. Specifically, we propose a set of candidate solutions from constrained optimal Markov line descriptors (COMLD) and then predict a candidate class using a classifier learned from training data. Our new framework will treat a multiple-line candidate as a new and independent TLS problem, and it will re-estimate method parameters to achieve better adaptations and thus better solutions. Our extensive experimental results indicate that this framework 1) reaches or outperforms state-of-the-art TLS solutions on various datasets; 2) is robust against various image variations including, but not limited to, rotation, noise, image, resolution, and text line spaces; and 3) is trainable for a new language/data and is self-adaptive to new testing data. This evidence suggests that our methodology is a highly promising solution to the TLS problem.

Recognition of old Greek document images containing polytonic (multi accent) characters is a challenging task due to the large number of existing character classes (more than 270) which cannot be handled sufficiently by current OCR technologies. Taking into account that the Greek polytonic system was used from the late antiquity until recently, a large amount of scanned Greek documents still remains without full test search capabilities. In order to assist the progress of relevant research, this paper introduces the first publicly available old Greek polytonic database GRPOLY-DB for the evaluation of several document image processing tasks. It contains both machine-printed and handwritten documents as well as annotation with ground-truth information that can be used for training and evaluation of the most commou document image processing tasks, i.e.. text line and word segmentation, test recognition, isolated character recognition and word spotting. Results using several representative baseline technologies are also presented in order to help researchers evaluate their methods and advance the frontiers of old Greek document image recognition and word spotting.

This paper presents the results of the Handwriting Segmentation Contest that was organized in the context of the ICDAR2013. The general objective of the contest was to use well established evaluation practices and procedures to record recent advances in off-line handwriting segmentation. Two benchmarking datasets, one for text line and one for word segmentation, were created in order to test and compare all submitted algorithms as well as some state-of-the-art methods for handwritten document image segmentation in realistic circumstances. Handwritten document images were produced by many writers in two Latin based languages (English and Greek) and in one Indian language (Bangla, the second most popular language in India). These images were manually annotated in order to produce the ground truth which corresponds to the correct text line and word segmentation results. The datasets of previously organized contests (ICDAR2007, ICDAR2009 and ICFHR2010 Handwriting Segmentation Contests) along with a dataset of Bangla document images were used as training dataset. Eleven methods are submitted in this competition. A brief description of the submitted algorithms, the evaluation criteria and the segmentation results obtained from the submitted methods are also provided in this manuscript.

Extraction of text lines from handwritten/printed document images is one of the important steps in the process of an Optical Character Recognition (OCR) system. In case of handwritten document images, presence of skewed, touching or overlapping text line(s) makes this process a real challenge to the researcher. In the present work, a new text line extraction technique based on line contour estimation is reported. Here, digitized document image is initially partitioned into a number of vertical fragments of equal width. Then all the line segments present in these vertical fragments are detected. Finally, the neighboring line segments are analyzed to place them inside the line boundary in which they actually belong. For experimental purpose, the developed technique is tested on CMATERdb1.2.1 database and present technique extracts 88.44% text lines successfully.