kaz29Вестник Казахстанско-Британского технического университетаHerald of the Kazakh-British Technical University1998-66882959-8109Казахстанско-Британский Технический Университет10.55452/1998-6688-2024-21-3-66-77kaz29-1370Research ArticleКОМПЬЮТЕРНЫЕ НАУКИCOMPUTER SCIENCEПРИМЕНЕНИЕ МЕТОДОВ МАШИННОГО ОБУЧЕНИЯ ДЛЯ ПОВЫШЕНИЯ УРОВНЯ ТОЧНОСТИ РЕЗУЛЬТАТОВ ОПТИЧЕСКОГО РАСПОЗНАВАНИЯ СИМВОЛОВAPPLICATION OF MACHINE LEARNING TECHNIQUES TO INCREASE THE LEVEL OF ACCURACY OF OPTICAL CHARACTER RECOGNITION RESULTShttps://orcid.org/0009-0007-1897-0768ВыходцеваВ. А.VykhodtsevaV. A.

магистрант 

070000, г. Усть-Каменогорск

Master student 

070000, Ust-Kamenogorsk

vykhodtseva.va@gmail.comhttps://orcid.org/0000-0002-6935-1066ПоповаГ. В.PopovaG. V.

к.ф–м.н., ассоциированный профессор 

070000, г. Усть-Каменогорск

PhD, Associate Professor 

070000, Ust-Kamenogorsk

gpopova@edu.ektu.kzКазахстанско-Американский свободный университетКазахстанKazakh-American Free UniversityKazakhstan2024011020242136677Copyright © Выходцева В.А., Попова Г.В., 20242024Выходцева В.А., Попова Г.В.Vykhodtseva V.A., Popova G.V.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://vestnik.kbtu.edu.kz/jour/article/view/1370

Одним из самых широко распространенных процессов современности, безусловно, является цифровизация, которая охватила все ключевые сферы жизни человечества. Развитие информационных технологий поспособствовало масштабным изменениям не только в повседневном аспекте жизнедеятельности, но и в более глобальном, автоматизировав сложные бизнес-процессы в сфере предпринимательства, экономики, здравоохранения. Переход к цифровым данным и документации обеспечил большую доступность необходимой информации, а также повысил эффективность ее анализа и обработки. В связи с данным фактом важное значение обрели технологии оптического распознавания символов (OCR), позволяющие определять и извлекать текстовые данные из изображений. OCR-технологии играют ключевую роль в цифровой трансформации общества, поскольку они исключают необходимость ручной работы с текстовой информацией на изображениях и применимы в автоматизации большинства бизнес-процессов, связанных с обработкой данных на бумажных носителях, например, при сборе статистических данных из бумажных форм, отражении бумажных документов в системе электронного документооборота, конвертации текстовой информации в аудиофайлы и так далее. Данная статья посвящена описанию технологии оптического распознавания символов, а также обзору методов машинного обучения, которые активно применяются в контексте ее современной реализации с целью улучшения качества получаемых результатов. Кроме того, в статье представлены принципы работы описываемых подходов, их возможности, а также некоторые ограничения, с которыми можно столкнуться при их использовании в тех или иных сценариях.

One of the most pervasive processes of modernity is undoubtedly digitalization, which has encompassed all key spheres of human life. The development of information technology has contributed to large-scale changes not only in the everyday aspect of life, but also more globally, automating complex business processes in the field of entrepreneurship, economics, and healthcare. The transition to digital data and documentation has provided greater accessibility to necessary information and has also enhanced the efficiency of its analysis and processing. Due to this fact, optical character recognition (OCR) technology has gained significant importance, enabling the identification and extraction of textual data from images. OCR systems play a pivotal role in the digital transformation of society as they eliminate the need for manual handling of textual information in images and are applicable in automating the majority of business processes associated with paper-based data processing, such as gathering statistical data from paper forms, reflecting paper documents in electronic document management systems, converting textual information into audio files, and so on. This paper is dedicated to describing optical character recognition technology, as well as providing an overview of machine learning techniques that are actively used in the context of its modern implementation, in order to enhance the quality of the obtained results. In addition, the paper presents the principles of operation of the described approaches, their capabilities, as well as some limitations that may be encountered when using them in various scenarios.

оптическое распознавание символовпризнакимашинное обучениеглубокое обучениесверточные нейронные сетирекуррентные нейронные сетиoptical character recognitionfeaturesmachine learningdeep learningconvolutional neural networkrecurrent neural networkReferencesSingh A., Bacchuwar K., Bhasin A. A survey of OCR Applications. International Journal of Machine Learning and Computing, 2012, vol. 2, no. 3, pp. 314–318.Singh A., Bacchuwar K., Bhasin A. A survey of OCR Applications. International Journal of Machine Learning and Computing, 2012, vol. 2, no. 3, pp. 314–318.Shruthi P. Verma D. C. A Detailed study and recent research on OCR. International Journal of Computer Science and Information Security, 2021, vol. 19, no. 2, pp. 52–66.Shruthi P. Verma D. C. A Detailed study and recent research on OCR. International Journal of Computer Science and Information Security, 2021, vol. 19, no. 2, pp. 52–66.Ahmed. M, Abidi A.I. Review on optical character recognition. International Research Journal of Engineering and Technology (IRJET), 2019, June, v0l. 06, issue 06, pp. 3666–3669.Ahmed. M, Abidi A.I. Review on optical character recognition. International Research Journal of Engineering and Technology (IRJET), 2019, June, v0l. 06, issue 06, pp. 3666–3669.Verma P., Foomani G. M. Improvement in OCR Technologies in Postal Industry Using CNN-RNN Architecture: Literature Review. International Journal of Machine Learning and Computing, 2022, vol. 12, no. 5, 154–163.Verma P., Foomani G. M. Improvement in OCR Technologies in Postal Industry Using CNN-RNN Architecture: Literature Review. International Journal of Machine Learning and Computing, 2022, vol. 12, no. 5, 154–163.What is optical character recognition (OCR)? Ibm.com. January 5, 2022. https://www.ibm.com/blog/optical-character-recognition/What is optical character recognition (OCR)? Ibm.com. January 5, 2022. https://www.ibm.com/blog/optical-character-recognition/Modi H., Parikh M. C. A Review on Optical Character Recognition Techniques International Journal of Computer Applications, 2017, vol. 160, no. 6, pp. 20–24.Modi H., Parikh M. C. A Review on Optical Character Recognition Techniques International Journal of Computer Applications, 2017, vol. 160, no. 6, pp. 20–24.Fateh A., Fateh M., Abolghasemi V. Enhancing optical character recognition: Efficient techniques for document layout analysis and text line detection. Engineering Reports, 2023, November, pp. 1–26.Fateh A., Fateh M., Abolghasemi V. Enhancing optical character recognition: Efficient techniques for document layout analysis and text line detection. Engineering Reports, 2023, November, pp. 1–26.Wang X.F., He Z.H., Wang K., Wang Y.F., Zou L. A survey of text detection and recognition algorithms based on deep learning technology. Neurocomputing, 2023, November, vol. 556.Wang X.F., He Z.H., Wang K., Wang Y.F., Zou L. A survey of text detection and recognition algorithms based on deep learning technology. Neurocomputing, 2023, November, vol. 556.AI vs. Machine Learning vs. DeepLearning vs. Neural Networks: What’s the difference? Ibm.com. July 6, 2023. https://www.ibm.com/blog/ai-vs-machine-learning-vs-deep-learning-vs-neural-networks/AI vs. Machine Learning vs. DeepLearning vs. Neural Networks: What’s the difference? Ibm.com. July 6, 2023. https://www.ibm.com/blog/ai-vs-machine-learning-vs-deep-learning-vs-neural-networks/Subedi B., Yunusov J., Gaybulayev A., Kim T. Development of a Low-cost Industrial OCR System with an End-to-end Deep Learning Technology. IEMEK J. Embed. Sys. Appl, 2020, April, pp. 51–60.Subedi B., Yunusov J., Gaybulayev A., Kim T. Development of a Low-cost Industrial OCR System with an End-to-end Deep Learning Technology. IEMEK J. Embed. Sys. Appl, 2020, April, pp. 51–60.Difference Between Machine Learning and Deep Learning. GeeksForGeeks.org. June 5, 2023. https://www.geeksforgeeks.org/difference-between-machine-learning-and-deep-learning/Difference Between Machine Learning and Deep Learning. GeeksForGeeks.org. June 5, 2023. https://www.geeksforgeeks.org/difference-between-machine-learning-and-deep-learning/Meng F., Ghena B. Research on Text Recognition Methods Based on Artificial Intelligence and Machine Learning. Advances in Computer and Communications, 2023, November, pp. 340–344.Meng F., Ghena B. Research on Text Recognition Methods Based on Artificial Intelligence and Machine Learning. Advances in Computer and Communications, 2023, November, pp. 340–344.Memon J., Sami M., Khan R. A. Handwritten Optical Character Recognition (OCR): A Comprehensive Systematic Literature Review (SLR). IEEE Access, 2020, July, pp. 142642–142668.Memon J., Sami M., Khan R. A. Handwritten Optical Character Recognition (OCR): A Comprehensive Systematic Literature Review (SLR). IEEE Access, 2020, July, pp. 142642–142668.Rakesh S., Reddy P.K., Prashanth V., Reddy K.S. Reddy. Handwritten text recognition using deep learning techniques: a survey. ICMED, 2024, pp. 1–8.Rakesh S., Reddy P.K., Prashanth V., Reddy K.S. Reddy. Handwritten text recognition using deep learning techniques: a survey. ICMED, 2024, pp. 1–8.Nikolenko S., Kadurin A., Arkhangelskaya E. Glyubokoye obucheniye: Seriya «Biblioteka programmista» (SPb.: Izd-vo Piter, 2018), pp. 231–259 [in Russian].Nikolenko S., Kadurin A., Arkhangelskaya E. Glyubokoye obucheniye: Seriya «Biblioteka programmista» (SPb.: Izd-vo Piter, 2018), pp. 231–259 [in Russian].Hemanth G.R., Jayasree M., Keerthi Venii S., Akshaya P., and R. Saranya. CNN-RNN based handwritten text recognition. ICTACT Journal on Soft Computing, 2021, October, vol. 12, pp. 2457–2463.Hemanth G.R., Jayasree M., Keerthi Venii S., Akshaya P., and R. Saranya. CNN-RNN based handwritten text recognition. ICTACT Journal on Soft Computing, 2021, October, vol. 12, pp. 2457–2463.Ahlawat S., Choudhary A., Nayyar A., Singh S., Yoon B. Improved Handwritten Digit Recognition Using Convolutional Neural Networks (CNN). Sensors, 2020, June, pp. 1–18.Ahlawat S., Choudhary A., Nayyar A., Singh S., Yoon B. Improved Handwritten Digit Recognition Using Convolutional Neural Networks (CNN). Sensors, 2020, June, pp. 1–18.Goodfellow I, Bengio Y, Courville A. Deep Learning, (The MIT Press, 2015), pp. 297–329.Goodfellow I, Bengio Y, Courville A. Deep Learning, (The MIT Press, 2015), pp. 297–329.Shinde S., Saraiya T. Using CRNN to Perform OCR over Forms. International Journal of Engineering Research & Technology (IJERT), vol. 9, pp. 319–323.Shinde S., Saraiya T. Using CRNN to Perform OCR over Forms. International Journal of Engineering Research & Technology (IJERT), vol. 9, pp. 319–323.Shi B., Bai X., & Yao C. An end-to-end trainable neural network for image-based sequence recognition and its application to scene text recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, pp. 2–3.Shi B., Bai X., & Yao C. An end-to-end trainable neural network for image-based sequence recognition and its application to scene text recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, pp. 2–3.Clausner C., Antonacopoulos A., Pletschacher S. Efficient and effective OCR engine training. International Journal on Document Analysis and Recognition (IJDAR), 2019, October, vol. 23, pp. 73–88.Clausner C., Antonacopoulos A., Pletschacher S. Efficient and effective OCR engine training. International Journal on Document Analysis and Recognition (IJDAR), 2019, October, vol. 23, pp. 73–88.Mursari L.R., Wibowo A. The Effectiveness of Image Preprocessing on Digital Handwritten Scripts Recognition with The Implementation of OCR Tesseract. Computer Engineering and Applications, 2021, October, vol. 10, pp. 177–186.Mursari L.R., Wibowo A. The Effectiveness of Image Preprocessing on Digital Handwritten Scripts Recognition with The Implementation of OCR Tesseract. Computer Engineering and Applications, 2021, October, vol. 10, pp. 177–186.Bagwe S., Shah V., Chauhan J., Harniya P., Tiwari A., Gupta V., Raikar D., Gada V., Bheda U., Mehta V., Warang M., Mehendale N. Optical character recognition using deep learning techniques for printed and handwritten documents. SSRN Electronic Journal, 2020, January, pp. 1–10.Bagwe S., Shah V., Chauhan J., Harniya P., Tiwari A., Gupta V., Raikar D., Gada V., Bheda U., Mehta V., Warang M., Mehendale N. Optical character recognition using deep learning techniques for printed and handwritten documents. SSRN Electronic Journal, 2020, January, pp. 1–10.Patel C., Patel A., Patel D. Optical Character Recognition by Open Source OCR Tool Tesseract: A Case Study. International Journal of Computer Applications, 2012, October, vol. 55, pp. 50–56.Patel C., Patel A., Patel D. Optical Character Recognition by Open Source OCR Tool Tesseract: A Case Study. International Journal of Computer Applications, 2012, October, vol. 55, pp. 50–56.Jain P., Dr. Taneja K., Dr. Taneja Harmunish. Which OCR toolset is good and why? A comparative study. Kuwait J.Sci., 2021, April, vol. 48 (2), pp. 1–12.Jain P., Dr. Taneja K., Dr. Taneja Harmunish. Which OCR toolset is good and why? A comparative study. Kuwait J.Sci., 2021, April, vol. 48 (2), pp. 1–12.

The authors declare that there are no conflicts of interest present.