Development of a learning program for basic digital image processing

Abstract

Background: Currently, radiography has widely shifted from the Film-Screen system to digital systems. This transition necessitates that radiological technologists possess a thorough understanding of digital image management and processing to produce high-quality medical radiographs. MATLAB, a program widely used for image processing, offers a Graphical User Interface (GUI) to facilitate faster and more intuitive use. However, its typical requirement for users to write commands in the Command Window can be confusing for those without a strong foundational knowledge of the program. Objective: This research aims to develop a program that assists in learning image processing commands and transformations, focusing on topics such as the Color Channels Model, Image Filtering, and Image Enhancement. Methods: The program was designed and developed specifically as a digital image processing learning tool. Its functionality was rigorously evaluated by comparing the images and histograms generated by the program with those produced by executing the same commands in MATLAB’s Workspace (Reference Code). To evaluate the program’s effectiveness in enhancing learning, pre- and post-test results were compared among 63 undergraduate students of Radiological Technology. Results: This comparison revealed that the Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) values were 0 for all commands, indicating identical results. Furthermore, the histograms from both images overlapped perfectly, confirming the accuracy and usability of the program's commands. Conclusion: The results showed that post-test scores were significantly higher than pre-test scores, with a statistically significant difference at a 95% confidence level.