Publication Search

Background: Automatic segmentation of the thyroid gland in ultrasonography (USG) images using deep learning requires a user-friendly interface to support diagnostic and educational processes. Purpose: This study aims to develop and implement a Graphical User Interface (GUI) that integrates a deep learning U-Net model for interactive and efficient segmentation and visualization of thyroid USG images. Method: The development method employed the Rapid Application Development (RAD) approach using MATLAB programming language. The GUI is designed to load transverse and sagittal USG images, display automatic segmentation results, and calculate thyroid gland volume based on dimensions measured automatically from the segmentation output. Testing was conducted using USG image data from 15 volunteers, and GUI functionality was evaluated using black box testing. Result: The GUI successfully displayed USG images and segmentation results with a responsive 4-panel interface; zoom, pan, and image navigation features functioned well. Automatic segmentation occurred in real-time after image input, and volume measurement results appeared automatically. Black box testing evaluation showed all GUI features operated as expected. The average Dice Similarity Coefficient (DSC) of 0.91 indicates high performance of the U-Net model in thyroid segmentation, consistent with previous findings. Statistical testing confirmed no significant difference between volume measurements using the application and manual methods (p = 0.953). Conclusion: This GUI implementation facilitates users in performing deep learning-based segmentation and visualization of thyroid USG images, improving efficiency and accuracy in thyroid volume measurement. The GUI has potential applications in clinical practice and radiology education.

Hydrogen-based hybrid microgrid systems have emerged as a promising solution to enhance renewable energy integration and improve energy supply reliability. By combining renewable sources such as solar and wind with hydrogen production and storage technologies, these systems address the intermittency of renewable power while ensuring continuous energy availability. This study evaluates the techno-economic feasibility, environmental impact, and scalability of hydrogen-based hybrid microgrids, with a focus on cost-effectiveness and system performance under varying operating conditions. Simulation tools, including HOMER Pro and MATLAB Simulink, are used to model the system and conduct sensitivity analyses on hydrogen production costs and demand fluctuations. Key performance indicators such as Levelized Cost of Energy (LCOE), Net Present Value (NPV), and CO₂ emissions reduction are assessed. The results show that although the system requires a high initial investment, it becomes economically viable over time due to reduced operational costs and improved efficiency. Additionally, the system demonstrates significant environmental benefits, outperforming conventional fossil fuel-based systems in terms of emissions reduction. Sensitivity analysis further indicates that advancements in hydrogen production technologies could substantially enhance economic feasibility. Overall, hydrogen-based hybrid microgrids offer a reliable and low-carbon energy solution, supporting sustainable energy transitions and reducing dependence on fossil fuels.

Background: Trigeminal neuralgia (TN) is often associated with neurovascular compression in the trigeminal nerve root entry zone, necessitating the simultaneous visualisation of nerves and blood vessels. Fusion of 3D SPACE and 3D TOF MRA images provides an integrated neurovascular view; however, not all hospitals have fusion software. This study developed a MATLAB-based image fusion method as an alternative and evaluated its equivalence to hospital-based fusion software.Methods: This study employed a descriptive quantitative research design, conducted in November 2025 at Diponegoro National Hospital and Dr Kariadi General Hospital in Semarang. A total of 16 brain MRI datasets (3D SPACE and 3D TOF MRA) were fused using hospital software and the MATLAB fusion application (MATLAB R2025b GUI). The fusion results were assessed by specialist radiologists. Diagnostic performance metrics (sensitivity, specificity, NDP, NDN, accuracy) were calculated, and paired differences were tested using the McNemar test. Intra-observer reliability was assessed using percentage agreement and Cohen’s Kappa.Results: MATLAB fusion yielded a sensitivity of 90.91%, specificity of 80.00%, NDP of 90.91%, NDN of 80.00%, and accuracy of 87.50%; the McNemar test (p=1.000) indicated no significant difference. Intra-observer reliability was very good (percent agreement 94%; Kappa 0.875). These findings indicate that MATLAB-based fusion is equivalent to hospital software fusion on the study data and has the potential to serve as an alternative in facilities without fusion software, provided that registration standardisation and user training are in place.

This study aims to develop an educational application based on a Graphical User Interface (GUI) using MATLAB App Designer that functions as an interactive simulation for evaluating blood pressure. The application allows users to input systolic and diastolic blood pressure values along with supporting information such as name and age. The input data are then analyzed and classified into several blood pressure categories according to the standards of the American Heart Association (AHA), including normal, hypotension, stage 1 hypertension, stage 2 hypertension, and hypertensive crisis. The classification results are presented visually through an interactive pie chart with dynamic percentages and legends to enhance user understanding. In addition, all data are automatically stored in a Microsoft Excel file containing a summary of blood pressure categories and session timestamps. The system is designed with a simple interface and intuitive interaction, making it suitable for early health education purposes. Although the application still relies on manual data input, it has the potential to serve as an effective learning tool for increasing public awareness of the importance of regular blood pressure monitoring.