Malaria detection using convolutional neural networks: A deep learning approach
Navin Kumar Agrawal · 2025
Malaria remains an important worldwide challenge, with millions of cases reported annually, especially in resource-constrained regions. Effective treatment and disease management of malaria depend on a prompt and precise diagnosis. In this work, we present a Convolutional Neural Network (CNN) based deep learning method for malaria identification. Our research leverages a comprehensive dataset comprising thousands of blood smear images collected from diverse malaria-endemic regions. We pre-process the images to enhance their quality and ensure consistency across the dataset. We then employ a CNN architecture, trained on this dataset, to automatically learn discriminative features from the blood smear images. The trained CNN model demonstrates remarkable performance in malaria detection, achieving an accuracy of over 95% on a hold-out test set. Moreover, our model exhibits high sensitivity and specificity, which are essential for reducing false positives and false negatives in the diagnosis of malaria. We compare our CNN-based approach with traditional machine learning methods and highlight its superior performance.