Optical character recognition (OCR) is one of the most popular techniques used for converting printed documents into machine-readable ones. While OCR engines can do well with modern text, their performance is unfortunately significantly reduced on historical materials. Additionally, many texts have already been processed by various out-of-date digitisation techniques. As a consequence, digitised texts are noisy and need to be post-corrected. This article clarifies the importance of enhancing quality of OCR results by studying their effects on information retrieval and natural language processing applications. We then define the post-OCR processing problem, illustrate its typical pipeline, and review the state-of-the-art post-OCR processing approaches. Evaluation metrics, accessible datasets, language resources, and useful toolkits are also reported. Furthermore, the work identifies the current trend and outlines some research directions of this field.

https://dl.acm.org/doi/pdf/10.1145/3453476

Survey of Post-OCR Processing Approaches

OCR post-processing is an important step for improving the quality of OCR output texts. Long short-term memory (LSTM) is a deep learning model, which has wide-range applications in many domains like time series prediction, natural language processing and speech recognition. In this paper, we propose an OCR error correction model using neural machine translation with bidirectional LSTM networks at syllable level. Vietnamese OCR text dataset for the model evaluation is outputted from an OCR engine based on the attention-based encoder-decoder (AED) model taking input of handwritten text in the benchmark database of the ICFHR 2018 Vietnamese online handwritten text recognition competition. The experimental results show that the proposed model helps decrease the word error rate in the OCR output texts of the above AED model by about 2%. The model performance is also discussed and compared to the other baseline methods in the competition.

OCR error correction for Vietnamese handwritten text using neural machine translation

Under the current artificial intelligence boom, machine translation is a research direction of natural language processing, which has important scientific research value and practical value. In practical applications, the variability of language, the limited capability of representing semantic information, and the scarcity of parallel corpus resources all constrain machine translation towards practicality and popularization. In this paper, we conduct deep mining of source language text data to express complex, high-level, and abstract semantic information using an appropriate text data representation model; then, for machine translation tasks with a large amount of parallel corpus, I use the capability of annotated datasets to build a more effective migration learning-based end-to-end neural network machine translation model on a supervised algorithm; then, for machine translation tasks with parallel corpus data resource-poor language machine translation tasks, migration learning techniques are used to prevent the overfitting problem of neural networks during training and to improve the generalization ability of end-to-end neural network machine translation models under low-resource conditions. Finally, for language translation tasks where the parallel corpus is extremely scarce but monolingual corpus is sufficient, the research focuses on unsupervised machine translation techniques, which will be a future research trend.

/pdf/a-study-on-the-intelligent-translation-model-for-english-2crqxdgt4s.pdf

A Study on the Intelligent Translation Model for English Incorporating Neural Network Migration Learning

OCR post-processing is an important step to improve OCR text accuracy. It includes two main tasks, error detection and error correction. Hill climbing algorithm is a heuristic search method used for solving optimization problems. In this paper, we present a novel OCR error correction approach using an adapted version of the Hill climbing algorithm. Correction candidates of OCR errors are explored by random character edits and evolved with the Hill climbing. The character edit patterns are obtained from the training data. The proposed model is evaluated on the benchmark dataset in the OCR post-correction competition of the International Conference on Document Analysis and Recognition 2017. It is shown that our model outperforms various baseline approaches in the competition. In addition, the randomness of the proposed algorithm is analyzed to verify its stability under parameter configurations.

Candidate word generation for OCR errors using optimization algorithm

Abstract The self-organizing migrating algorithm (SOMA) is a population-based meta-heuristic that belongs to swarm intelligence. In the last 20 years, we can observe two main streams in the publications. First, novel approaches contributing to the improvement of its performance. Second, solving the various optimization problems. Despite the different approaches and applications, there exists no work summarizing them. Therefore, this work reviews the research papers dealing with the principles and application of the SOMA. The second goal of this work is to provide additional information about the performance of the SOMA. This work presents the comparison of the selected algorithms. The experimental results indicate that the best-performing SOMAs provide competitive results comparing the recently published algorithms. 

/pdf/self-organizing-migrating-algorithm-review-improvements-and-27noj70b.pdf

Self-organizing migrating algorithm: review, improvements and comparison

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.

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

OCR post-processing is an essential step to improve the accuracy of OCR-generated texts by detecting and correcting OCR errors. In this paper, the OCR texts are resulted from an OCR engine which is based on the attention-based encoder-decoder model for unconstrained Vietnamese handwriting. We identify various kinds of Vietnamese OCR errors and their possible causes. Detailed statistics of Vietnamese OCR errors are provided and analyzed at both character level and syllable level, using typical OCR error characteristics such as error rate, error mapping/edit, frequency and error length. Furthermore, the statistical analyses are done on training and test sets of a benchmark database to infer whether the test set is the appropriate representative of the training set regarding the OCR error characteristics. We also discuss the choice of designing OCR post-processing approaches at character level or at syllable level relying on provided statistics of studied datasets.

An In-depth Analysis of OCR Errors for Unconstrained Vietnamese Handwriting.

Candidate word generation by character edit operations is an important method that has been employed in many OCR error correction approaches. In this paper, we study how character edit distances impact the performance of OCR error correction. We propose the algorithm of generating correction candidates with different edit distances. Correction candidates for both non-word and real-word errors are considered. The candidates are scored and ranked based on linguistic features and edit probability. The experiments are tested on the VNOnDB database used in the Vietnamese online handwritten text recognition competition (VOHTR 2018). We evaluate the error correction performance on different edit distances in terms of two error metrics, character error rate (CER) and word error rate (WER). It is shown that the edit distances of 1 and 2 obtain better correction results instead of higher edit distances. 

OCR Error Correction for Vietnamese OCR Text with Different Edit Distances

Different types of OCR errors often occur in OCR texts due to the low quality of scanned document images or limitations in OCR software. In this paper, we propose a novel unsupervised approach for OCR error correction. Correction candidates for OCR errors are generated and explored in their neighborhoods using correction character edits controlled by an adapted hill-climbing algorithm. Correction characters are extracted from only original ground truth texts, which do not depend on OCR texts in training data. A weighted objective function used to score and rank correction candidates is heuristically tested to find optimal weight combinations. The proposed model is evaluated on an OCR text dataset originating from the Vietnamese handwritten database in the ICFHR 2018 Vietnamese online handwritten text recognition competition. The proposed model is also verified concerning its stability and complexity. The experimental results show that our model achieves competitive performance compared to the other models in the ICFHR 2018 competition.

Nguyet-Minh Phan

Papers

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

An In-depth Analysis of OCR Errors for Unconstrained Vietnamese Handwriting.

OCR Error Correction for Vietnamese OCR Text with Different Edit Distances

An Efficient Unsupervised Approach for OCR Error Correction of Vietnamese OCR Text