Original article
Optical Character Recognition-Based Data Entry Program for Reporting Radiation Workers’ Health Examinations
Yun-Min Woo, Assistant1 and Joo-Young Lee, Professor2
1Department of Health & Medical Field, Shingu College
2Department of Radiological Science, Shingu College
Correspondence to Joo-Young Lee, PhD. Department of Radiological Science, Shingu College. 377 Gwangmyeong-ro, Jungwon-gu, Seongnam, 13174, Republic of Korea. Tel: +82-31-740-1576, E-mail: ljy0706@shingu.ac.kr
Volume 30, Number 2, Article 22, November 2026. Korean J Nucl Med Technol 2026;30(2):22. https://doi.org/10.12972/kjnmt.2026.30.2.22
Received on October 08, 2026, Revised on October 26, 2026, Accepted on October 29, 2026, Published on November 30, 2026.
This is an Open Access article distributed under the terms of the Creative Commons CC BY 4.0 license (https://creativecommons.org/licenses/by/4.0/) which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
Purpose: This study developed a data entry program that extracts reporting items from health examination forms of radiation workers processed by optical character recognition, and evaluated its accuracy and work time. Materials and Methods: The program extracts 18 required items of the national reporting form and flags uncertain values for staff review. Forms of 100 virtual workers were scanned and processed by optical character recognition. Cell-level accuracy of 16 items read from the forms was compared with an answer key, and work time for 100 workers was estimated with the keystroke-level model. Results: Accuracy was 86.0% (95% Confidence Interval, 84.2–87.6%), but only 49.4% in descriptive cells with written content. Error flags identified 61.2–61.6% of errors. With full review, estimated work time decreased from 146.4 to 63.8 minutes (56.4%). Conclusion: The program halved work time but should be used with full review by staff.
Keywords
Radiation Workers, Health Examination, Optical Character Recognition, Data Entry, Keystroke-Level Model