Optical Character Recognition Applications Research Articles

Implementing AI on low-power embedded devices for digital water meter identification and data transfer via lora network

This study introduces an artificial intelligence system implemented on the ESP32-CAM platform, aimed at conducting optical character recognition (OCR) on water meters. Leveraging LoRa technology for data transmission ensures efficient energy utilization and convenient long-range communication capabilities. The system achieves an impressive accuracy rate of 98.2% in identifying water meter readings, showcasing its reliability. Proposed as a feasible solution, it offers the advantages of low energy consumption, cost-effectiveness, and flexibility in widespread deployment, particularly leveraging existing water meter infrastructure. Thus, this research presents a promising choice for various applications beyond merely reading water meter readings. Its efficient and accurate OCR functionality makes it suitable for diverse scenarios, ranging from smart city initiatives to industrial automation processes. With its ability to integrate seamlessly into existing infrastructure and deliver reliable results, this system stands poised to revolutionize OCR applications in various domains, contributing to enhanced efficiency and productivity.

PERANCANGAN SISTEM WAREHOUSE BERBASIS TEKNOLOGI OCR UNTUK MENINGKATKAN EFEKTIVITAS DAN EFISIENSI

PERANCANGAN SISTEM WAREHOUSE BERBASIS TEKNOLOGI OCR UNTUK MENINGKATKAN EFEKTIVITAS DAN EFISIENSI

Candidate Authentication using OCR Techniques

This paper explores the application of Optical Character Recognition (OCR) techniques for authenticating candidates based on identity documents. As advancements in technology continue to redefine various industries, the integration of OCR into candidate authentication processes offers a streamlined and efficient solution. The aim is to design a system which gets the image of the identity proof and the details are being retrieved using the character segmentation which is done by a feature extraction optical character recognition algorithm (OCR). The authentication process encompasses document validation and implementation of security measures to safeguard against fraud.By combining YOLO object detection, Optical Character Recognition (OCR) and fuzzy matching technique ,this approach aims to enhance the accuracy and reliability of candidate authentication. This underscores the importance of a comprehensive authentication framework, incorporating both technological and procedural elements, to establish a robust and secure candidate authentication process in diverse contexts. Keywords: ID Verification,Optical Character Recognition (OCR), Text Extraction, YOLO,Object Detection,Image Segmentation, EasyOCR , Fuzzy Matching.

Design and development of OCR software for remote measurement and calibration

The advent of digital technology has led to significant changes in work processes in many organizations and industries worldwide including Thailand. This work aims to apply digital technology to improve measurement and metrology in terms of accuracy, productivity, and variability. We developed machine vision software for monitoring and collecting digital data from measurement systems. The application of Optical Character Recognition (OCR) technology in metrology is the first step toward digitalization, enabling enterprises to improve their measurement processes by automating data collection, analysis, and management. Moreover, it can be integrated with machine learning algorithms to enhance the automation of complex measurement tasks. To achieve this goal, an OCR software for data recording was designed in the LabVIEW environment. The image captured by the camera is processed by the LabVIEW OCR system which is trained continuously until it can recognize characters consistently and accurately. This software has been tested with the calibration systems which has a seven-segment digital format. The result shows that the software can acquire and transmit the instrument reading with reliable and exceptional accuracy. In addition, it can also reduce the steps and time required for measurement. In conclusion, the application of OCR technology can improve the efficiency, productivity, and accuracy of measurement. It also has a promising potential to enable remote measurement and calibration according to ISO/IEC 17025 possible.

Minimalist DCT-based Depthwise Separable Convolutional Neural Network Approach for Tangut Script

The Tangut script, a lesser-explored dead script comprising numerous characters, has received limited attention in deep learning research, particularly in the field of optical character recognition (OCR). Existing OCR studies primarily focus on widely-used characters like Chinese characters and employ deep convolutional neural networks (CNNs) or combinations with recurrent neural networks (RNNs) to enhance accuracy in character recognition. In contrast, this study takes a counterintuitive approach to develop an OCR model specifically for the Tangut script. We utilize shorter layers with slimmer filters using a depthwise separable convolutional neural network (DSCNN) architecture. Furthermore, we preprocess the dataset using a frequency-based transformation, namely the Discrete Cosine Transform (DCT). The results demonstrate successful training of the model, showcasing faster convergence and higher accuracy compared to traditional deep neural networks commonly used in OCR applications.

An optimized deep learning model for optical character recognition applications

The convolutional neural networks (CNN) are among the most utilized neural networks in various applications, including deep learning. In recent years, the continuing extension of CNN into increasingly complicated domains has made its training process more difficult. Thus, researchers adopted optimized hybrid algorithms to address this problem. In this work, a novel chaotic black hole algorithm-based approach was created for the training of CNN to optimize its performance via avoidance of entrapment in the local minima. The logistic chaotic map was used to initialize the population instead of using the uniform distribution. The proposed training algorithm was developed based on a specific benchmark problem for optical character recognition applications; the proposed method was evaluated for performance in terms of computational accuracy, convergence analysis, and cost.

Artificially Intelligent Readers: An Adaptive Framework for Original Handwritten Numerical Digits Recognition with OCR Methods

Advanced artificial intelligence (AI) techniques have led to significant developments in optical character recognition (OCR) technologies. OCR applications, using AI techniques for transforming images of typed text, handwritten text, or other forms of text into machine-encoded text, provide a fair degree of accuracy for general text. However, even after decades of intensive research, creating OCR with human-like abilities has remained evasive. One of the challenges has been that OCR models trained on general text do not perform well on localized or personalized handwritten text due to differences in the writing style of alphabets and digits. This study aims to discuss the steps needed to create an adaptive framework for OCR models, with the intent of exploring a reasonable method to customize an OCR solution for a unique dataset of English language numerical digits were developed for this study. We develop a digit recognizer by training our model on the MNIST dataset with a convolutional neural network and contrast it with multiple models trained on combinations of the MNIST and custom digits. Using our methods, we observed results comparable with the baseline and provided recommendations for improving OCR accuracy for localized or personalized handwritten text. This study also provides an alternative perspective to generating data using conventional methods, which can serve as a gold standard for custom data augmentation to help address the challenges of scarce data and data imbalance.

A Novel Short-Memory Sequence-Based Model for Variable-Length Reading Recognition of Multi-Type Digital Instruments in Industrial Scenarios

As a practical application of Optical Character Recognition (OCR) for the digital situation, the digital instrument recognition is significant to achieve automatic information management in real-industrial scenarios. However, different from the normal digital recognition task such as license plate recognition, CAPTCHA recognition and handwritten digit recognition, the recognition task of multi-type digital instruments faces greater challenges due to the reading strings are variable-length with different fonts, different spacing and aspect ratios. In order to overcome this shortcoming, we propose a novel short-memory sequence-based model for variable-length reading recognition. First, we involve shortcut connection strategy into traditional convolutional structure to form a feature extractor for capturing effective features from characters with different fonts of multi-type digital instruments images. Then, we apply an RNN-based sequence module, which strengthens short-distance dependencies while reducing the long-distance trending memory of the reading string, to greatly improve the robustness and generalization of the model for invisible data. Finally, a novel short-memory sequence-based model consisting of a feature extractor, an RNN-based sequence module and the CTC, is proposed for variable-length reading recognition of multi-type digital instruments. Experimental results show that this method is effective on variable-length instrument reading recognition task, especially for invisible data, which proves that our method has outstanding generalization and robustness in real-industrial applications.

Automated Vehicle Number Plate Detection and Recognition

Abstract: In this project, a Digital Image Processing-based prototype is developed. Actions such as Image Acquisition, enhancement that is pre-processing, Segmentation of the license plate and then application of OCR (Optical Character Recognition) is applied to store the number on text form. The plate number is displayed as text on the terminal using the principle of OCR with help of Tesseract engine. It is seen that the security forces and authorities face problems whenever security forces chase a vehicle or they can’t catch a vehicle which broke traffic rules. Authorities find it very hectic on a busy day to log the vehicle numbers manually in a parking lot.

Application of Software Methods for Automated Processing of Sources of Personal Origin

The subject of this research is software methods of automated preprocessing of historical sources and the development of effective solutions to problems when working with sources of personal origin. The article analyzes the current situation in the use of modern software methods. The authors demonstrate the main range of arguments for which such historical sources from a technical point of view should be considered separately. A methodological analysis of the features of the application of optical character recognition based on preprocessed data is carried out. Special attention is paid to the advantages and key parameters of the effectiveness of the final result of work when using automated text processing, including the further use of OCR methods. The scientific novelty of the research lies in the proposal and detailed description of a software solution to the current problem based on machine learning methods. The developed program has three phases of working with digital copies of sources of personal origin. It is based on the use of the OpenCV library and solving a number of problems using the Hough transform. Based on the general analysis of the study, we can highlight the main advantages of automated preprocessing of scanned documents: reducing time, improving accuracy, combating distortion and optimizing the process. The presented results of successful testing of the developed solution allow us to judge the possible areas of its effective application.

Balinese Script Recognition Using Tesseract Mobile Framework

One of the main factors causing the decline in the use of Balinese Script is that Balinese people are less interested in reading Balinese Script because of their reluctance to learn Balinese Script, which is relatively complicated in the recognition process. The development of computer technology has now been used to help by performing character recognition or known as Optical Character Recognition (OCR). Developing the OCR application for Balinese Script is an effort to help preserve, from the technology side, as a means of education related to Balinese Script. In this study, that development was conducted by using a Tesseract OCR engine that consists of several stages, i.e., the first one is to prepare the dataset, the second one is to generate the dataset using the Web Scraping method, the third one is to train the OCR engine using the generated dataset, and finally, the fourth one is to implement the generated language model into a mobile-based application. The study results prove that the dataset generation process using the Web Scraping method can be a better choice when faced with a training dataset that requires a large dataset compared to several previous studies of non-Latin character recognition. In those studies, the jTessBox tools were used, which took time because they had to select per character for a dataset. The best result of the language model is a combination of character, word, sentence, and paragraph datasets (hierarchical combination of character, word, sentence, and paragraph datasets) with a coincidence rate of 66.67%. The more diverse and structured hierarchical datasets used, the higher the coincidence rate.

Digitizing historical balance sheet data: A practitioner’s guide

This paper discusses how to successfully digitize large-scale historical micro-data by augmenting optical character recognition (OCR) engines with pre- and post-processing methods. Although OCR software has improved dramatically in recent years due to improvements in machine learning, off-the-shelf OCR applications still present high error rates which limit their applications for accurate extraction of structured information. Complementing OCR with additional methods can however dramatically increase its success rate, making it a powerful and cost-efficient tool for economic historians. This paper showcases these methods and explains why they are useful. We apply them against two large balance sheet datasets and introduce quipucamayoc, a Python package containing these methods in a unified framework.

Penerapan Metode Long Short-Term Memory Pada Pendataan Warga Berbasis Android

Along with technological developments, utilization in data processing both input and output become very important. Starting from inputting data manually, it proves that the use of technology is still not fully used. Data that has been processed conventionally sometimes encounters problems caused by human error. Not only that, when data must be entered into a computer, it requires additional time and when natural disasters such as floods, earthquakes, landslides occur. Census files that are still in the conventional form are likely to be damaged and will take time to re-data. This is the main reason for RT 004 RW 008 to need an efficient data collection system. The application of Optical Character Recognition (OCR) technology in this study is expected to be a solution in increasing the efficiency of the data input process. In this study, the input process is carried out in real time using the Long Short-Term Memory (LSTM) method to scan data on e-KTP. The results of testing OCR technology with the LSTM method in this study for the three experiments carried out were 1) The e-KTP detection experiment resulted in an accuracy of 100%, 2) The trial of reading 19 attributes on the e-KTP resulted in an accuracy of 98.42%, 3) Testing the error reading of the e-KTP attribute resulted in an accuracy of 93.56%. Overall, the accuracy of the LSTM method in word detection on e-KTP results in an accuracy of 92.1%. These results indicate that the application of OCR technology with the LSTM method is one of the right solutions for the data input process.

Image Classification for Edge-Cloud Setting: A Comparison Study for OCR Application

The increasing number of smart devices has led to a rise in the complexity and volume of the image generated. Deep learning is an increasingly common approach for image classification, a fundamental task in many applications. Due to its high computational requirements, implementation in edge devices becomes challenging. Cloud computing serves as an enabler, allowing devices with limited resources to perform deep learning. For cloud computing, however, latency is an issue and is undesirable. Edge computing addresses the issue by redistributing data and tasks closer to the edge. Still, a suitable offloading strategy is required to ensure optimal performance with methods such as LeNet-5, OAHR, and Autoencoder (ANC) as feature extractors paired with different classifiers (such as artificial neural network (ANN) and support vector machine (SVM)). In this study, models are evaluated using a dataset representing Optical Character Recognition (OCR) task. The OCR application has recently been used in many task-offloading studies. The evaluation is based on the time performance and scoring criteria. In terms of time performance, a fully connected ANN using features from the ANC is faster by a factor of over 60 times compared to the fastest performing SVM. Moreover, scoring performance shows that the SVM is less prone to overfit in the case of a noisy or imbalanced dataset in comparison with ANN. So, adopting SVM in which the data distribution is unspecified will be wiser as there is a lower tendency to overfit. The training and inference time, however, are generally higher than ANN.

First study of the RuPatient health information system with optical character recognition of medical records based on machine learning

RuPatient health information system (HIS) is a computer program consisting of a doctor-patient web user interface, which includes algorithms for recognizing medical record text and entering it into the corresponding fields of the system.Aim. To evaluate the effectiveness of RuPatient HIS in actual clinical practice.Material and methods. The study involved 10 cardiologists and intensivists of the department of cardiology and сardiovascular intensive care unit of the L. A. Vorokhobov City Clinical Hospital 67 We analyzed images (scanned copies, photos) of discharge reports from patients admitted to the relevant departments in 2021. The following fields of medical documentation was recognized: Name, Complaints, Anamnesis of life and illness, Examination, Recommendations. The correctness and accuracy of recognition of entered information were analyzed. We compared the recognition quality of RuPatient HIS and a popular optical character recognition application (FineReader for Mac).Results. The study included 77 pages of discharge reports of patients from various hospitals in Russia from 50 patients (men, 52%). The mean age of patients was 57,7±7,9 years. The number of reports with correctly recognized fields in various categories using the program algorithms was distributed as follows: Name — 14 (28%), Diagnosis — 13 (26%), Complaints — 40 (80%), Anamnesis — 14 (28%), Examination — 24 (48%), Recommendations — 46 (92%). Data that was not included in the category was also recognized and entered in the comments field. The number of recognized words was 549±174,9 vs 522,4±215,6 (p=0,5), critical errors in words — 2,1±1,6 vs 4,4±2,8 (p<0,001), non-critical errors — 10,3±4,3 vs 5,6±3,3 (p<0,001) for RuPatient HIS and optical character recognition application for a personal computer, respectively.Conclusion. The developed RuPatient HIS, which includes a module for recognizing medical records and entering data into the corresponding fields, significantly increases the document management efficiency with high quality of optical character recognition based on neural network technologies and the automation of filling process.

WITHDRAWN: Web based optical character recognition application using flask and tesseract

Optical Character Recognition (OCR) is a methodology utilized to transform over checked pictures into editable content configuration. OCR is basically the mechanical or electronic interpretation of manually written pictures or typewritten content into machine-editable content. Many antithetic kinds of Optical Character Recognition (OCR) implementations are commercially accessible today and it is a tract of investigation in pattern recognition, man-made reasoning and machine visual sense. In the proposed work images processed independently (.png, .jpg, and .gif files). Manually written acknowledgement is utilized regularly to depict the capacity of a PC to translate human composition into content. This may take in one of the two different ways, either by examining written content or by composing straightforwardly on fringe input gadgets. This work also assembles something comparative utilizing an tool of Optical Character Recognition (OCR) available in Google Tesseract OCR Engine alongside Python, Open-CV to specific fictional characters present in pictures. Primarily, in this proposed work a website is being set up using a web framework called Flask which is written in Python. Finally, the objective of this to deploy a web based Optical Character Recognition (OCR) using Python and Python based web tool i.e. Flask and the Tesseract engine. The communication protocol thus used is HTTP. The maximum accuracy obtained by authors is 94% for handwritten black & white images and 87% for handwritten color images. Authors continued their research for digitized images and the results obtained are 98% for black and white and 93% color images.

Application of optical character recognition with natural language processing for large-scale quality metric data extraction in colonoscopy reports

Background and AimsColonoscopy is commonly performed for colorectal cancer screening in the United States. Reports are often generated in a non-standardized format and are not always integrated into electronic health records. Thus, this information is not readily available for streamlining quality management, participating in endoscopy registries, or reporting of patient- and center-specific risk factors predictive of outcomes. We aim to demonstrate the use of a new hybrid approach using natural language processing of charts that have been elucidated with optical character recognition processing (OCR/NLP hybrid) to obtain relevant clinical information from scanned colonoscopy and pathology reports, a technology co-developed by Cleveland Clinic and eHealth Technologies (West Henrietta, NY, USA). MethodsThis was a retrospective study conducted at Cleveland Clinic, Cleveland, Ohio, and the University of Minnesota, Minneapolis, Minnesota. A randomly sampled list of outpatient screening colonoscopy procedures and pathology reports was selected. Desired variables were then collected. Two researchers first manually reviewed the reports for the desired variables. Then, the OCR/NLP algorithm was used to obtain the same variables from 3 electronic health records in use at our institution: Epic (Verona, Wisc, USA), ProVation (Minneapolis, Minn, USA) used for endoscopy reporting, and Sunquest PowerPath (Tucson, Ariz, USA) used for pathology reporting. ResultsCompared with manual data extraction, the accuracy of the hybrid OCR/NLP approach to detect polyps was 95.8%, adenomas 98.5%, sessile serrated polyps 99.3%, advanced adenomas 98%, inadequate bowel preparation 98.4%, and failed cecal intubation 99%. Comparison of the dataset collected via NLP alone with that collected using the hybrid OCR/NLP approach showed that the accuracy for almost all variables was >99%. ConclusionsOur study is the first to validate the use of a unique hybrid OCR/NLP technology to extract desired variables from scanned procedure and pathology reports contained in image format with an accuracy >95%.

Efficient Hardware Architectures for 1D- and MD-LSTM Networks

Recurrent Neural Networks, in particular One-dimensional and Multidimensional Long Short-Term Memory (1D-LSTM and MD-LSTM) have achieved state-of-the-art classification accuracy in many applications such as machine translation, image caption generation, handwritten text recognition, medical imaging and many more. However, high classification accuracy comes at high compute, storage, and memory bandwidth requirements, which make their deployment challenging, especially for energy-constrained platforms such as portable devices. In comparison to CNNs, not so many investigations exist on efficient hardware implementations for 1D-LSTM especially under energy constraints, and there is no research publication on hardware architecture for MD-LSTM. In this article, we present two novel architectures for LSTM inference: a hardware architecture for MD-LSTM, and a DRAM-based Processing-in-Memory (DRAM-PIM) hardware architecture for 1D-LSTM. We present for the first time a hardware architecture for MD-LSTM, and show a trade-off analysis for accuracy and hardware cost for various precisions. We implement the new architecture as an FPGA-based accelerator that outperforms NVIDIA K80 GPU implementation in terms of runtime by up to 84× and energy efficiency by up to 1238× for a challenging dataset for historical document image binarization from DIBCO 2017 contest, and a well known MNIST dataset for handwritten digits recognition. Our accelerator demonstrates highest accuracy and comparable throughput in comparison to state-of-the-art FPGA-based implementations of multilayer perceptron for MNIST dataset. Furthermore, we present a new DRAM-PIM architecture for 1D-LSTM targeting energy efficient compute platforms such as portable devices. The DRAM-PIM architecture integrates the computation units in a close proximity to the DRAM cells in order to maximize the data parallelism and energy efficiency. The proposed DRAM-PIM design is 16.19 × more energy efficient as compared to FPGA implementation. The total chip area overhead of this design is 18 % compared to a commodity 8 Gb DRAM chip. Our experiments show that the DRAM-PIM implementation delivers a throughput of 1309.16 GOp/s for an optical character recognition application.

XtremDew: a platform for cooperative tasks and data schedulers

With the emergence of big data, data scheduling is becoming an important field of research in distributed computing. Software data scheduler often relies on data management policies that can be defined by the user and provide high level features. Such advanced features become necessary nowadays to execute data intensive applications, and this implies that data and task schedulers should cooperate closely to address the large data processing issue and ensure an optimal distribution of data intensive applications. In this paper, we propose XtremDew, the data and task cooperative scheduler platform. We deal with the distribution of the optical character recognition (OCR) on large scale. We show, in particular, the benefit of the focus on data scheduling to distribute our OCR application. We build the data driven distributing platform by combining two existing middleware: BitDew, as the data scheduler, and XtremWeb-HEP, as the task scheduler. Taking advantage of both middlewares, XtremDew provides new features. To evaluate the efficiency of our approach, we compare different strategies of scheduling tasks and data and we present several scenarios that illustrate the benefits of using XtremDew to execute data-intensive applications.

Image Spam Filtering using Machine Learning Techniques

Unsolicited visual data is undesirable in any form. The art of hiding malicious content in images and adding them as attachments to electronic mails has become a popular nuisance. In recent years, attackers have developed various new techniques to evade traditional spam classification systems. Text-based spam classification has been in focus for a long time and, researchers have successfully created a prodigal system for identifying spam text in electronic mails using Optical Character Recognition technology. In the last decade, extensive work has been performed to tackle image spam but with unsatisfactory results. Various algorithms and data augmentation techniques are used today to develop an optimal model for image spam recognition. Many of these proposed systems come close to the ideal system but do not provide 100 percent accuracy. This paper highlights the role of three popular techniques in image spam filtering. We discuss the importance and application of Optical Character Recognition, Support Vector Machines and, Artificial Neural Networks in unsolicited visual data filtering. This paper sheds light on the algorithms of these techniques. We provide a comparison of their accuracy, which helps us draw useful insights for developing a robust unsolicited visual data classification system. This paper aims to bring clarity regarding the feasibility of using these techniques to develop an unsolicited visual data filtering system. This paper records that the most favourable results are obtained using Artificial Neural Networks.