Optical Character Recognition Errors Research Articles

Toward a Period-specific Optimized Neural Network for OCR Error Correction of Historical Hebrew Texts

Over the past few decades, large archives of paper-based historical documents, such as books and newspapers, have been digitized using the Optical Character Recognition (OCR) technology. Unfortunately, this broadly used technology is error-prone, especially when an OCRed document was written hundreds of years ago. Neural networks have shown great success in solving various text processing tasks, including OCR post-correction. The main disadvantage of using neural networks for historical corpora is the lack of sufficiently large training datasets they require to learn from, especially for morphologically rich languages like Hebrew. Moreover, it is not clear what are the optimal structure and values of hyperparameters (predefined parameters) of neural networks for OCR error correction in Hebrew due to its unique features. Furthermore, languages change across genres and periods. These changes may affect the accuracy of OCR post-correction neural network models. To overcome these challenges, we developed a new multi-phase method for generating artificial training datasets with OCR errors and hyperparameters’ optimization for building an effective neural network for OCR post-correction in Hebrew. To evaluate the proposed approach, a series of experiments using several literary Hebrew corpora from various periods and genres were conducted. The obtained results demonstrate that (1) training a network on texts from a similar period dramatically improves the network's ability to fix OCR errors, (2) using the proposed error injection algorithm, based on character-level period-specific errors, minimizes the need for manually corrected data and improves the network accuracy by 9%, (3) the optimized network design improves the accuracy by 3% compared to the state-of-the-art network, and (4) the constructed optimized network outperforms neural machine translation models and industry-leading spellcheckers. The proposed methodology may have practical implications for digital humanities projects that aim to search and analyze OCRed documents in Hebrew and potentially other morphologically rich languages.

PNRank: Unsupervised ranking of person name entities from noisy OCR text

Text databases have grown tremendously in number, size, and volume over the last few decades. Optical Character Recognition (OCR) software is used to scan the text and make them available in online repositories. The OCR transcription process is often not accurate resulting in large volumes of garbled text in the repositories. Spell correction and other post-processing of OCR text often prove to be very expensive and time-consuming. While it is possible to rely on the OCR model to assess the quality of text in a corpus, many natural language processing and information retrieval tasks prefer the extrinsic evaluation of the effect of noise on the task at hand. This paper examines the effect of noise on the unsupervised ranking of person name entities by first populating a list of person names using an out-of-the-box Named Entity Recognition (NER) software, extracting content-based features for the identified entities, and ranking them using a novel unsupervised Kernel Density Estimation (KDE) based ranking algorithm. This generative model has the ability to learn rankings using the data distribution and therefore requires limited manual intervention. Empirical results are presented on a carefully curated parallel corpus of OCR and clean text and “in the wild” using a large real-world corpus. Experiments on the parallel corpus reveals that even with a reasonable degree of noise in the dataset, it is possible to generate ranked lists using the KDE algorithm with a high degree of precision and recall. Furthermore, since the KDE algorithm has comparable performance to state-of-the-art unsupervised rankers, using it on real-world corpora is feasible. The paper concludes by reflecting on other methods for enhancing the performance of the unsupervised algorithm on OCR text such as cleaning entity names, disambiguating names concatenated to one another and correcting OCR errors that are statistically significant in the corpus.

Warped Document Image Correction Based on Checkboard Pattern and Geometric Transformation

Abstract— Document image warping problem refers to the process of geometrically transforming 2D images. In this work we aim to solve the warped document image problems. The idea of the proposed method is inspired from the concept of image registration to align two images (distorted document and undistorted ground truth checkboard). In this paper, we propose document image dewarping method based on checkboard calibration pattern and geometric transformation. Some pairs of control points are selected between the two images to define the mapping function between them. The dewarping process transforms the warped document image according to the geometric transformation defined by the calculated mapping function. Results on document dewarping dataset CBDAR 2007 demonstrate the success of our strategy. OCR (Optical Character Recognition) error metrics are also used to gauge the success of the suggested approach. Based on the quantitative and qualitative metrics the proposed method outperforms the compared state of the art method.

OCR error correction using correction patterns and self-organizing migrating algorithm

Optical character recognition (OCR) systems help to digitize paper-based historical achieves. However, poor quality of scanned documents and limitations of text recognition techniques result in different kinds of errors in OCR outputs. Post-processing is an essential step in improving the output quality of OCR systems by detecting and cleaning the errors. In this paper, we present an automatic model consisting of both error detection and error correction phases for OCR post-processing. We propose a novel approach of OCR post-processing error correction using correction pattern edits and evolutionary algorithm which has been mainly used for solving optimization problems. Our model adopts a variant of the self-organizing migrating algorithm along with a fitness function based on modifications of important linguistic features. We illustrate how to construct the table of correction pattern edits involving all types of edit operations and being directly learned from the training dataset. Through efficient settings of the algorithm parameters, our model can be performed with high-quality candidate generation and error correction. The experimental results show that our proposed approach outperforms various baseline approaches as evaluated on the benchmark dataset of ICDAR 2017 Post-OCR text correction competition.

Organizing an in-class hackathon to correct PDF-to-text conversion errors of Genomics & Informatics 1.0.

This paper describes a community effort to improve earlier versions of the full-text corpus of Genomics & Informatics by semi-automatically detecting and correcting PDF-to-text conversion errors and optical character recognition errors during the first hackathon of Genomics & Informatics Annotation Hackathon (GIAH) event. Extracting text from multi-column biomedical documents such as Genomics & Informatics is known to be notoriously difficult. The hackathon was piloted as part of a coding competition of the ELTEC College of Engineering at Ewha Womans University in order to enable researchers and students to create or annotate their own versions of the Genomics & Informatics corpus, to gain and create knowledge about corpus linguistics, and simultaneously to acquire tangible and transferable skills. The proposed projects during the hackathon harness an internal database containing different versions of the corpus and annotations.

Generating high-quality data abstractions from scanned clinical records: text-mining-assisted extraction of endometrial carcinoma pathology features as proof of principle

ObjectiveMedical research studies often rely on the manual collection of data from scanned typewritten clinical records, which can be laborious, time consuming and error prone because of the need to...

Flexible character accuracy measure for reading-order-independent evaluation

The extraction of textual information from scanned document pages is a fundamental stage in any digitisation effort and directly determines the success of the overall document analysis and understanding application scenarios. To evaluate and improve the performance of optical character recognition (OCR), it is necessary to measure the accuracy of that step alone, without the influence of the processing steps that precede it (e.g. text block segmentation and ordering). Current OCR performance evaluation measures (based on edit distance) are strongly subjective as they need to first serialise the entire text in the documents – a process influenced heavily by the specific reading order determined (often wrongly, especially in cases of multicolumn and complex layouts) by processing steps prior to OCR. This paper presents a new objective and practical edit-distance-based character recognition accuracy measure which overcomes those limitations. It achieves its independence from the reading order by comparing sub-strings of text in a flexible way (i.e. allowing for ordering variations). The precision of the flexible character accuracy measure enables the effective tuning of complete digitisation workflows (as OCR errors are isolated and other steps can be evaluated and optimised separately). For the same reason, it also enables a better estimation of post-OCR (manual) correction effort required. The proposed character accuracy measure has been systematically analysed and validated under lab conditions as well as successfully used in practice in a number of high-profile international competitions since 2017.

Toward the optimized crowdsourcing strategy for OCR post-correction

PurposeDigitization of historical documents is a challenging task in many digital humanities projects. A popular approach for digitization is to scan the documents into images, and then convert images into text using optical character recognition (OCR) algorithms. However, the outcome of OCR processing of historical documents is usually inaccurate and requires post-processing error correction. The purpose of this paper is to investigate how crowdsourcing can be utilized to correct OCR errors in historical text collections, and which crowdsourcing methodology is the most effective in different scenarios and for various research objectives.Design/methodology/approachA series of experiments with different micro-task’s structures and text lengths were conducted with 753 workers on the Amazon’s Mechanical Turk platform. The workers had to fix OCR errors in a selected historical text. To analyze the results, new accuracy and efficiency measures were devised.FindingsThe analysis suggests that in terms of accuracy, the optimal text length is medium (paragraph-size) and the optimal structure of the experiment is two phase with a scanned image. In terms of efficiency, the best results were obtained when using longer text in the single-stage structure with no image.Practical implicationsThe study provides practical recommendations to researchers on how to build the optimal crowdsourcing task for OCR post-correction. The developed methodology can also be utilized to create golden standard historical texts for automatic OCR post-correction.Originality/valueThis is the first attempt to systematically investigate the influence of various factors on crowdsourcing-based OCR post-correction and propose an optimal strategy for this process.

A novel Arabic OCR post-processing using rule-based and word context techniques

Optical character recognition (OCR) is the process of recognizing characters automatically from scanned documents for editing, indexing, searching, and reducing the storage space. The resulted text from the OCR usually does not match the text in the original document. In order to minimize the number of incorrect words in the obtained text, OCR post-processing approaches can be used. Correcting OCR errors is more complicated when we are dealing with the Arabic language because of its complexity such as connected letters, different letters may have the same shape, and the same letter may have different forms. This paper provides a statistical Arabic language model and post-processing techniques based on hybridizing the error model approach with the context approach. The proposed model is language independent and non-constrained with the string length. To the best of our knowledge, this is the first end-to-end OCR post-processing model that is applied to the Arabic language. In order to train the proposed model, we build Arabic OCR context database which contains 9000 images of Arabic text. Also, the evaluation of the OCR post-processing system results is automated using our novel alignment technique which is called fast automatic hashing text alignment. Our experimental results show that the rule-based system improves the word error rate from 24.02% to become 20.26% by using a training data set of 1000 images. On the other hand, after this training, we apply the rule-based system on 500 images as a testing dataset and the word error rate is improved from 14.95% to become 14.53%. The proposed hybrid OCR post-processing system improves the results based on using 1000 training images from a word error rate of 24.02% to become 18.96%. After training the hybrid system, we used 500 images for testing and the results show that the word error rate enhanced from 14.95 to become 14.42. The obtained results show that the proposed hybrid system outperforms the rule-based system.

Improving Information Retrieval Performance on OCRed Text in the Absence of Clean Text Ground Truth

OCR errors in text harm information retrieval performance. Much research has been reported on modelling and correction of Optical Character Recognition (OCR) errors. Most of the prior work employ language dependent resources or training texts in studying the nature of errors. However, not much research has been reported that focuses on improving retrieval performance from erroneous text in the absence of training data. We propose a novel approach for detecting OCR errors and improving retrieval performance from the erroneous corpus in a situation where training samples are not available to model errors. In this paper we propose a method that automatically identifies erroneous term variants in the noisy corpus, which are used for query expansion, in the absence of clean text. We employ an effective combination of contextual information and string matching techniques. Our proposed approach automatically identifies the erroneous variants of query terms and consequently leads to improvement in retrieval performance through query expansion. Our proposed approach does not use any training data or any language specific resources like thesaurus for identification of error variants. It also does not expend any knowledge about the language except that the word delimiter is blank space. We have tested our approach on erroneous Bangla (Bengali in English) and Hindi FIRE collections, and also on TREC Legal IIT CDIP and TREC 5 Confusion track English corpora. Our proposed approach has achieved statistically significant improvements over the state-of-the-art baselines on most of the datasets.

Can Morphological Analyzers Improve the Quality of Optical Character Recognition?

Optical Character Recognition (OCR) can substantially improve the usability of digitized documents. Language modeling using word lists is known to improve OCR quality for English. For morphologically rich languages, however, even large word lists do not reach high coverage on unseen text. Morphological analyzers offer a more sophisticated approach, which is useful in many language processing applications. is paper investigates language modeling in the open-source OCR engine Tesseract using morphological analyzers. We present experiments on two Uralic languages Finnish and Erzya. According to our experiments, word lists may still be superior to morphological analyzers in OCR even for languages with rich morphology. Our error analysis indicates that morphological analyzers can cause a large amount of real word OCR errors.

Daoulagad — A Celto-Slavic OCR Dictionary

In this paper we present Daoulagad [dɔwˈlaːgat], a mobile Celtic-Russian dictionary, supporting Optical Character Recognition (OCR). The dictionary provides Cymraeg↔Русский, Cymraeg↔English, Cymraeg↔Gaeilge, Cymraeg↔Brezhoneg, Gaeilge↔Русский, Gaeilge↔English, Brezhoneg↔Русский, English↔Русский translations, supports initial consonant mutations and ‘Item and Arrangement’ & ‘Word and Paradigm’ morphological models. OCR capabilities make it possible to use iPhone or Android phone's camera as input device. OCR errors are corrected using trigram frequencies calculated over an extensive corpus. Also supported are Belarusian, Bulgarian, Croatian/Serbian, Czech, Polish, Slovak, Slovenian and Ukranian (as well as English, French, German, Italian, Latin, Portuguese, Spanish, Thai, Arabic and hanzi/kanji).

Annotated chemical patent corpus: a gold standard for text mining.

Exploring the chemical and biological space covered by patent applications is crucial in early-stage medicinal chemistry activities. Patent analysis can provide understanding of compound prior art, novelty checking, validation of biological assays, and identification of new starting points for chemical exploration. Extracting chemical and biological entities from patents through manual extraction by expert curators can take substantial amount of time and resources. Text mining methods can help to ease this process. To validate the performance of such methods, a manually annotated patent corpus is essential. In this study we have produced a large gold standard chemical patent corpus. We developed annotation guidelines and selected 200 full patents from the World Intellectual Property Organization, United States Patent and Trademark Office, and European Patent Office. The patents were pre-annotated automatically and made available to four independent annotator groups each consisting of two to ten annotators. The annotators marked chemicals in different subclasses, diseases, targets, and modes of action. Spelling mistakes and spurious line break due to optical character recognition errors were also annotated. A subset of 47 patents was annotated by at least three annotator groups, from which harmonized annotations and inter-annotator agreement scores were derived. One group annotated the full set. The patent corpus includes 400,125 annotations for the full set and 36,537 annotations for the harmonized set. All patents and annotated entities are publicly available at www.biosemantics.org.

De-identification of health records using Anonym: effectiveness and robustness across datasets.

Evaluate the effectiveness and robustness of Anonym, a tool for de-identifying free-text health records based on conditional random fields classifiers informed by linguistic and lexical features, as well as features extracted by pattern matching techniques. De-identification of personal health information in electronic health records is essential for the sharing and secondary usage of clinical data. De-identification tools that adapt to different sources of clinical data are attractive as they would require minimal intervention to guarantee high effectiveness. The effectiveness and robustness of Anonym are evaluated across multiple datasets, including the widely adopted Integrating Biology and the Bedside (i2b2) dataset, used for evaluation in a de-identification challenge. The datasets used here vary in type of health records, source of data, and their quality, with one of the datasets containing optical character recognition errors. Anonym identifies and removes up to 96.6% of personal health identifiers (recall) with a precision of up to 98.2% on the i2b2 dataset, outperforming the best system proposed in the i2b2 challenge. The effectiveness of Anonym across datasets is found to depend on the amount of information available for training. Findings show that Anonym compares to the best approach from the 2006 i2b2 shared task. It is easy to retrain Anonym with new datasets; if retrained, the system is robust to variations of training size, data type and quality in presence of sufficient training data.

Natural Language Processing for Historical Texts Michael Piotrowski (Leibniz Institute of European History) Morgan & Claypool (Synthesis Lectures on Human Language Technologies, edited by Graeme Hirst, volume 17), 2012, ix+157 pp; paperbound, ISBN 978-1608459469

<b>Natural Language Processing for Historical Texts</b> <b>Michael Piotrowski</b> (Leibniz Institute of European History) Morgan &amp; Claypool (Synthesis Lectures on Human Language Technologies, edited by Graeme Hirst, volume 17), 2012, ix+157 pp; paperbound, ISBN 978-1608459469

Optical character recognition errors and their effects on natural language processing

Errors are unavoidable in advanced computer vision applications such as optical character recognition, and the noise induced by these errors presents a serious challenge to downstream processes that attempt to make use of such data. In this paper, we apply a new paradigm we have proposed for measuring the impact of recognition errors on the stages of a standard text analysis pipeline: sentence boundary detection, tokenization, and part-of-speech tagging. Our methodology formulates error classification as an optimization problem solvable using a hierarchical dynamic programming approach. Errors and their cascading effects are isolated and analyzed as they travel through the pipeline. We present experimental results based on a large collection of scanned pages to study the varying impact depending on the nature of the error and the character(s) involved. This dataset has also been made available online to encourage future investigations.

Document cleanup using page frame detection

When a page of a book is scanned or photocopied, textual noise (extraneous symbols from the neighboring page) and/or non-textual noise (black borders, speckles, ...) appear along the border of the document. Existing document analysis methods can handle non-textual noise reasonably well, whereas textual noise still presents a major issue for document analysis systems. Textual noise may result in undesired text in optical character recognition (OCR) output that needs to be removed afterwards. Existing document cleanup methods try to explicitly detect and remove marginal noise. This paper presents a new perspective for document image cleanup by detecting the page frame of the document. The goal of page frame detection is to find the actual page contents area, ignoring marginal noise along the page border. We use a geometric matching algorithm to find the optimal page frame of structured documents (journal articles, books, magazines) by exploiting their text alignment property. We evaluate the algorithm on the UW-III database. The results show that the error rates are below 4% each of the performance measures used. Further tests were run on a dataset of magazine pages and on a set of camera captured document images. To demonstrate the benefits of using page frame detection in practical applications, we choose OCR and layout-based document image retrieval as sample applications. Experiments using a commercial OCR system show that by removing characters outside the computed page frame, the OCR error rate is reduced from 4.3 to 1.7% on the UW-III dataset. The use of page frame detection in layout-based document image retrieval application decreases the retrieval error rates by 30%.

Selecting a restoration technique to minimize OCR error

This paper introduces a learning problem related to the task of converting printed documents to ASCII text files. The goal of the learning procedure is to produce a function that maps documents to restoration techniques in such a way that on average the restored documents have minimum optical character recognition error. We derive a general form for the optimal function and use it to motivate the development of a nonparametric method based on nearest neighbors. We also develop a direct method of solution based on empirical error minimization for which we prove a finite sample bound on estimation error that is independent of distribution. We show that this empirical error minimization problem is an extension of the empirical optimization problem for traditional M-class classification with general loss function and prove computational hardness for this problem. We then derive a simple iterative algorithm called generalized multiclass ratchet (GMR) and prove that it produces an optimal function asymptotically (with probability 1). To obtain the GMR algorithm we introduce a new data map that extends Kesler's construction for the multiclass problem and then apply an algorithm called Ratchet to this mapped data, where Ratchet is a modification of the Pocket algorithm . Finally, we apply these methods to a collection of documents and report on the experimental results.

TELLTALE: Experiments in a dynamic hypertext environment for degraded and multilingual data

Methods and tools for finding documents relevant to a user's needs in document corpora can be found in the information retrieval, library science, and hypertext communities. Typically, these systems provide retrieval capabilities for fairly static corpora, their algorithms are dependent on the language for which they are written, e.g. English, and they do not perform well when presented with misspelled words or text that has been degraded by OCR (optical character recognition) techniques. In this article, we present experimentation results for the TELLTALE system. TELLTALE is a dynamic hypertext environment that provides full-text search from a hypertext-style user interface for text corpora that may be garbled by OCR or transmission errors, and that may contain languages other than English. TELLTALE uses several techniques based on n-grams (n character sequences of text). With these results we show that the dynamic linkage mechanisms in TELLTALE are tolerant of garbles in up to 30% of the characters in the body of the text. © 1996 John Wiley & Sons, Inc.

Validation of image defect models for optical character recognition

Considers the problem of evaluating character image generators that model distortions encountered in optical character recognition (OCR). While a number of such defect models have been proposed, the contention that they produce the desired result is typically argued in an ad hoc and informal way. The authors introduce a rigorous and more pragmatic definition of when a model is accurate: they say a defect model is validated if the OCR errors induced by the model are indistinguishable from the errors encountered when using real scanned documents. The authors describe four measures to quantify this similarity, and compare and contrast them using over ten million scanned and synthesized characters in three fonts. The measures differentiate effectively between different fonts and different scans of the same font regardless of the underlying text.