Binary Segmentation of Malaria Parasites Using U-Net Segmentation Approach: A Case of Rwanda

Abstract

Abstract Malaria is a significant health issue in Rwanda. Its accurate identification is essential for effective treatment. Traditional methods, such as microscopy, often face limitations in these contexts. This paper investigates how advanced machine learning techniques can address diagnostic challenges commonly encountered in resource-limited settings like Rwanda. A powerful deep learning framework known as U-Net was utilized in this study to identify different types of malaria. This method demonstrated the ability to accurately identify the disease at a highly detailed level, yielding promising results. The findings from this study could contribute to the development of computer-aided diagnostic tools specifically designed for regions with limited resources. These tools could assist healthcare professionals in decision-making processes and enhance patient outcomes.