Malaria Cell Detection Using Deep Neural Networks

TL;DR

Malaria diagnosis remains a critical global health challenge, especially in low-resource settings where microscopy-based assessment is labor-intensive and operator-dependent. The paper presents a custom ResNet50-like deep learning architecture with residual and bottleneck blocks, trained via transfer learning and data augmentation on the Kaggle Malaria Cell Images Dataset (27,558 images) to automatically distinguish parasitized from uninfected cells. It reports strong performance on the test set: accuracy $97.8\%$, precision $98.6\%$, recall $97.2\%$, and F1-score $97.8\%$, with training accuracy around $99\%$ and validation accuracy around $98\%$, and includes a Streamlit-based web application to enable rapid, accessible predictions. The work demonstrates the practical potential of deep learning for automated malaria detection and provides a deployable tool that could support healthcare providers in resource-limited settings, while outlining clear directions for dataset expansion, model optimization, real-time deployment, and explainability.

Abstract

Malaria remains one of the most pressing public health concerns globally, causing significant morbidity and mortality, especially in sub-Saharan Africa. Rapid and accurate diagnosis is crucial for effective treatment and disease management. Traditional diagnostic methods, such as microscopic examination of blood smears, are labor-intensive and require significant expertise, which may not be readily available in resource-limited settings. This project aims to automate the detection of malaria-infected cells using a deep learning approach. We employed a convolutional neural network (CNN) based on the ResNet50 architecture, leveraging transfer learning to enhance performance. The Malaria Cell Images Dataset from Kaggle, containing 27,558 images categorized into infected and uninfected cells, was used for training and evaluation. Our model demonstrated high accuracy, precision, and recall, indicating its potential as a reliable tool for assisting in malaria diagnosis. Additionally, a web application was developed using Streamlit to allow users to upload cell images and receive predictions about malaria infection, making the technology accessible and user-friendly. This paper provides a comprehensive overview of the methodology, experiments, and results, highlighting the effectiveness of deep learning in medical image analysis.

Figures (5)

• Figure 1: Training and Validation Accuracy
• Figure 2: Training and Validation Loss
• Figure 3: Confusion Matrix
• Figure 4: Web Application - Image Upload Interface
• Figure 5: Web Application - Classification Result