Detection Of Malaria Infected Red Blood Cells Using Optimized Machine Learning Technique

Abstract

Malaria continues to pose a significant public health challenge, particularly in Sub-Saharan Africa where it accounts for millions of deaths annually. Traditional microscopic examination of blood smears, while being the gold standard, is labor-intensive and heavily dependent on expert interpretation. This study proposes an automated detection system for malaria-infected red blood cells using Convolutional Neural Network (CNN), a deep learning approach that has demonstrated remarkable success in medical image analysis. The research utilized a publicly available malaria dataset from GitHub containing 27,558 cell images. A GUI-based application was developed to allow users to upload and analyze blood cell images in real-time. The CNN model was designed with multiple convolutional layers, pooling layers, and fully connected layers to automatically extract relevant features from blood cell images and classify them as either infected or uninfected. The dataset underwent extensive preprocessing including image normalization and augmentation to enhance model performance. The proposed CNN model achieved an accuracy of 96.47%, precision of 95.89%, recall of 96.84%, and F1-score of 96.36%, significantly outperforming traditional machine learning algorithms such as Support Vector Machines (91.23%), Random Forest (92.78%), and Artificial Neural Networks (93.45%) reported in previous studies. The results demonstrate that deep learning approaches, particularly CNNs, offer a more efficient and accurate alternative for automated malaria detection. The developed GUI application provides a practical, accessible system for malaria diagnosis assistance, potentially reducing the burden on healthcare professionals and enabling faster diagnosis in resource-limited settings.