SYSTEM FOR RECOGNIZING AND CLASSIFYING FINANCIAL DOCUMENTS
DOI:
https://doi.org/10.34132/mspc2025.01.14.13Keywords:
document processing automation, optical character recognition (OCR), semantic text analysis, digital transformation of enterprises, modular system architecture.Abstract
Annotation: This study focuses on the automation of document processing, specifically the extraction of requisites from waybills (TTN) to facilitate digital transformation in enterprises. The main objective is to develop a flexible software system capable of accurately extracting data from documents of various formats. The system incorporates image preprocessing techniques, optical character recognition (OCR) using Nicomsoft OCR, and a template-based semantic text analysis approach. Extracted data undergoes normalization to ensure standardized storage. The implementation is based on Java, utilizing Swing for the user interface, Apache POI for document processing, and Firebird SQL for data storage. The modular architecture enables seamless integration with corporate platforms and scalability for processing different document types. By enhancing document management efficiency, this solution supports enterprise digitalization, optimizing data processing speed and accuracy while reducing operational costs. The adaptability of the system makes it a valuable tool for modernizing administrative workflows across various industries.
References
Shahar Markovitch and Paul Willmott/McKinsey – Accelerating the digitalization of business processes[Електронний ресурс] - Режим доступу: https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/accelerating-the-digitization-of-business-processes
Sayallar S., Sayar A., Babalik N.: An OCR Engine for Printed Receipt Images using Deep Learning Techniques. International Journal of Advanced Computer Science and Applications, Vol. 14, No. 2, 2023. URL: https://avesis.kocaeli.edu.tr/yayin/aa68b0f4-568a-46da-b680-98d60f7e561c/an-ocr-engine-for-printed-receipt-images-using-deep-learning-techniques (дата звернення: 22.04.2025)
Yue A. Automated Receipt Image Identification, Cropping, and Parsing, 2018. URL: https://www.cs.princeton.edu/courses/archive/spring18/cos598B/public/projects/System/COS598B_spr2018_ReceiptParsing.pdf (дата звернення: 23.04.2025)
Drobac S., Linden K.: Optical character recognition with neural networks and post-correction with finite state methods. International Journal on Document Analysis and Recognition (IJDAR) (2020) 23:279–295 URL: https://link.springer.com/article/10.1007/s10032-020-00359-9 (дата звернення: 23.04.2025)
