2024
An OCR-based application using Tesseract engine to extract text information from ultrasound B-MODE images
BLAHUTA, Jiří; Tomáš SOUKUP; Jiří KOZEL and Jiří MARTINŮBasic information
Original name
An OCR-based application using Tesseract engine to extract text information from ultrasound B-MODE images
Authors
BLAHUTA, Jiří; Tomáš SOUKUP; Jiří KOZEL and Jiří MARTINŮ
Edition
3792. vyd. Neuveden, Proceedings of the 24th Conference Information Technologies – Applications and Theory (ITAT 2024), p. 105-110, 6 pp. 2024
Publisher
CEUR-WS
Other information
Language
English
Type of outcome
Proceedings paper
Field of Study
10201 Computer sciences, information science, bioinformatics
Country of publisher
Slovakia
Confidentiality degree
is not subject to a state or trade secret
Publication form
electronic version available online
References:
Marked to be transferred to RIV
Yes
RIV identification code
RIV/47813059:19240/24:A0001431
Organization unit
Faculty of Philosophy and Science in Opava
ISSN
EID Scopus
Keywords in English
C# Tesseract Engine; Image Text Extraction; Medical OCR Application; Ultrasound B-MODE OCR
Tags
International impact, Reviewed
Changed: 29/1/2025 14:40, Mgr. Kamil Matula, Ph.D.
Abstract
In the original language
In this paper we introduce our developed OCR-based application focused on the extraction of text data from ultrasound B-MODE images. These images contain not only visual image but also additional, important text information about the examination, image data, etc. The extraction of these data is helpful in clinical practice. The application has a simple, user-friendly interface to use. The core of the algorithm to extract the text data is focused on Tesseract engine with C# programming language; front-end is a Windows Forms application. Although this software is fast, simple and user-friendly, in some cases, the recognition could produce a error, like patient’s name incorrectly recognized and/or some characters are missing or mistaken. That means, some text in the image could be missed or misinterpreted in comparison with the original. However, while the software could be a helpful tool, the recognition accuracy should be improved. After that, the application can be used for different types of medical images, e.g. CT/CTA, PET, SPECT and many more. Currently achieved accuracy is about 90 % in average. The authors discuss some ideas to increase the accuracy of the recognition and also some front-end features that can be improved for more comfortable use. Extracted text data can be saved as a CSV file for further processing.