A Lightweight Ensemble Deep Learning Framework for Automated Malaria Parasite Detection from Microscopic Blood Cell Images

We have proposed a lightweight deep learning framework for automated malaria parasite detection from microscopic blood cell images, aiming to provide accurate and computationally efficient diagnosis.

We have introduced a weighted ensemble learning strategy by combining Custom CNN and DenseNet121 to improve classification performance, robustness, and generalization capability.

We have validated the effectiveness of the proposed framework using multiple evaluation metrics, including accuracy, precision, recall, F1-score, ROC-AUC, PR-AUC, confidence interval analysis, and 5-fold cross-validation.