AP25794129 – Development of an algorithm for filtering and pre-processing biomedical images for cardiac diagnostics

Objective of the project – to apply machine learning methods to automated filtering and pre-processing of biomedical images, which will improve the accuracy and efficiency of cardiovascular disease diagnostics.

Relevance: The relevance of this project is driven by the high prevalence of cardiovascular diseases and the need to improve diagnostic accuracy. The quality of biomedical images directly affects the correctness of diagnosis, while noise and low clarity often complicate analysis. This increases the workload of medical professionals and raises the risk of diagnostic errors. Modern machine learning methods provide opportunities for automated image filtering and enhancement. The implementation of such technologies improves the efficiency of early diagnosis and accelerates medical data processing. Therefore, the project is highly relevant as it contributes to improving the quality of healthcare and reducing health risks for patients.

Scientific supervisor: Master of Technical Sciences, Senior lecturer, Kabdullin Maxat Amangeldiuly

Expected and achieved results: Within the project, an expanded dataset of biomedical images was created, enriched across different classes, providing a reliable foundation for training and testing algorithms. Data preprocessing was carried out, including the removal of duplicates, noise reduction, and image normalization. These measures significantly improved data quality and ensured the proper functioning of subsequent analytical methods. A prototype algorithm for biomedical image filtering was developed based on fundamental digital processing methods. The algorithm is capable of identifying signs of cardiovascular diseases in images. The prototype was tested, confirming its effectiveness in improving image quality and reducing noise levels. As a result of the work, a technological foundation for improving diagnostic accuracy was established. The scientific novelty of the project lies in adapting filtering and preprocessing methods to Kazakhstani biomedical data. This approach allows consideration of local characteristics and increases the reliability of diagnostic results. Overall, the obtained results confirm the successful completion of this project stage and its practical significance.

List of publications with links to them

1. Kabdullin M., Naizabayeva L., Kabdullin A., Zhonkeshova A.
   Development of an algorithm for automatic analysis of biomedical images in cardiology using machine learning methods and prospects for further analysis // Вестник Академии гражданской авиации (АГА). – Алматы, 2025. – В печати.
2. Kabdullin M., Kairbekov A., Kabdullin A., Naizabayeva L., Mukhit A.
   Design and Implementation of a Web-Based Prototype for Cardiological Image Management // DTESI 2025: 10th International Conference on Digital Technologies in Education, Science and Industry, November 19–20, 2025, Almaty, Kazakhstan. – Алматы, 2025. – В печати.