Automated Malaria Ring Form Classification in Blood Smear Images Using Ensemble Parallel Neural Networks

TL;DR

This paper introduces an AI system that can detect malaria ring-form parasites in blood smear images with high accuracy, potentially improving malaria diagnosis efficiency.

Contribution

The novel contribution is a dual-branch neural network architecture combining convolutional and attention-based features for malaria ring-form classification.

Findings

The model achieved an ROC–AUC of approximately 0.99 and a macro F1-score of 0.97 on an independent test set.

Moderate-capacity feature fusion outperformed more complex architectures, which suffered from higher false positives.

The system shows promise for automated malaria screening but requires external validation before clinical use.

Abstract

Manual microscopy for malaria diagnosis is labor-intensive and prone to inter-observer variability. This study presents an automated binary classification approach for detecting malaria ring-form infections in thin blood smear single-cell images using a parallel neural network framework. Utilizing a balanced Kaggle dataset of 27,558 erythrocyte crops, images were standardized to 128 × 128 pixels and subjected to on-the-fly augmentation. The proposed architecture employs a dual-branch fusion strategy, integrating a convolutional neural network for local morphological feature extraction with a multi-head self-attention branch to capture global spatial relationships. Performance was rigorously evaluated using 10-fold stratified cross-validation and an independent 10% hold-out test set. Results demonstrated high-level discrimination, with all models achieving an ROC–AUC of approximately 0.99. The primary model (Model#1) attained a peak mean accuracy of 0.9567 during cross-validation and 0.97 accuracy (macro F1-score: 0.97) on the independent test set. In contrast, increasing architectural complexity in Model#3 led to a performance decline (0.95 accuracy) due to higher false-positive rates. These findings suggest that moderate-capacity feature fusion, combining convolutional descriptors with attention-based aggregation, provides a robust and generalizable solution for automated malaria screening without the risks associated with over-parameterization. Despite a strong performance, immediate clinical use remains limited because the model was developed on pre-segmented single-cell images, and external validation is still required before routine implementation.

Figures