Hierarchical Deep Learning for Malaria Diagnosis: Integrating YOLO-Based Detection and Uncertainty-Aware Cell Classification

• Luciano Luz Beylouni Farias UFRGS
• Mateus Balda Mota UFRGS
• Karin Becker UFRGS
• Mariana Recamonde-Mendoza UFRGS

DOI: https://doi.org/10.5753/sbcas.2026.21624

Resumo

Malaria diagnosis through optical microscopy of blood smears remains the gold standard but is labor-intensive and subject to inter-observer variability. This study proposes a hierarchical deep learning pipeline that integrates cell detection in microscope field of view (FoV) images with classification of individual red blood cells. YOLOv8 is used to detect candidate cells and the extracted regions are classified by convolutional neural networks with uncertainty quantification and Grad-CAM explainability. MobileNetV2 achieved 95.17% accuracy in the classification of isolated cells with efficient prediction sets. However, under domain shift conditions, conformal prediction revealed a drop in reliability (coverage from 95% to approximately 86%), indicating reduced generalization. Grad-CAM analysis also revealed the Clever Hans effect, where models relied on image artifacts rather than biological features. These results highlight the importance of integrating reliability and explainability mechanisms for computer-aided malaria diagnosis.

Referências

Ahamed, M. F., Nahiduzzaman, M., Mahmud, G., Shafi, F. B., Ayari, M. A., Khandakar, A., Abdullah-Al-Wadud, M., and Islam, S. R. (2025). Improving malaria diagnosis through interpretable customized cnns architectures. Scientific Reports, 15(1):6484.

Angelopoulos, A. N. and Bates, S. (2021). A gentle introduction to conformal prediction and distribution-free uncertainty quantification. arXiv preprint arXiv:2107.07511.

Asif, S., Khan, S. U. R., Zheng, X., and Zhao, M. (2024). MozzieNet: A deep learning approach to efficiently detect malaria parasites in blood smear images. International Journal of Imaging Systems and Technology, 34(1):e22953.

Boström, H. (2022). crepes: a python package for generating conformal regressors and predictive systems. In Conformal and Probabilistic Prediction with Applications, pages 24–41. PMLR.

Doi, K. (2007). Computer-aided diagnosis in medical imaging: historical review, current status and future potential. Computerized medical imaging and graphics, 31(4-5):198–211.

He, K., Zhang, X., Ren, S., and Sun, J. (2016). Identity mappings in deep residual networks. In European Conference on Computer Vision (ECCV), pages 630–645. Springer.

Kingma, D. P. and Ba, J. (2015). Adam: A method for stochastic optimization. In International Conference on Learning Representations (ICLR).

Ljosa, V., Sokolnicki, K. L., and Carpenter, A. E. (2012). Annotated high-throughput microscopy image sets for validation. Nature Methods, 9(7):637.

Otesteanu, C. F., Caldelari, R., Heussler, V., and Sznitman, R. (2024). Machine learning for predicting plasmodium liver stage development in vitro using microscopy imaging. Computational and Structural Biotechnology Journal, 24:334–342.

Rajaraman, S., Antani, S. K., Poostchi, M., Silamut, K., Hossain, M. A., Maude, R. J., Jaeger, S., and Thoma, G. R. (2018). Pre-trained convolutional neural networks as feature extractors toward improved malaria parasite detection in thin blood smear images. PeerJ, 6:e4568.

Ramos, J. d. S., Vieira, I. H. P., Rocha, W. S., Esquerdo, R. P., Watanabe, C. Y. V., and Zanchi, F. B. (2024). A transfer learning approach to identify plasmodium in microscopic images. PLoS Computational Niology, 20(8):e1012327.

Ramos-Briceño, D. A., Flammia-D’Aleo, A., Fernández-López, G., Carrión-Nessi, F. S., and Forero-Peña, D. A. (2025). Deep learning-based malaria parasite detection: convolutional neural networks model for accurate species identification of plasmodium falciparum and plasmodium vivax. Scientific Reports, 15(1):3746.

Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., and Chen, L.-C. (2018). Mobilenetv2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 4510–4520.

Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., and Batra, D. (2017). Grad-cam: Visual explanations from deep networks via gradient-based localization. In Proceedings of the IEEE international conference on computer vision, pages 618–626.

Tangpukdee, N., Duangdee, C., Wilairatana, P., and Krudsood, S. (2009). Malaria diagnosis: a brief review. The Korean journal of parasitology, 47(2):93.

Voß, Y., Klaus, S., Guizetti, J., and Ganter, M. (2023). Plasmodium schizogony, a chronology of the parasite’s cell cycle in the blood stage. PLoS Pathogens, 19(3):e1011157.

Vásquez-Venegas, C., Wu, C., Sundar, S., Prôa, R., Beloy, F. J., Medina, J. R., McNichol, M., Parvataneni, K., Kurtzman, N., Mirshawka, F., Aguirre-Jerez, M., Ebner, D. K., and Celi, L. A. (2025). Detecting and mitigating the clever hans effect in medical imaging: A scoping review. J Imaging Inform Med.

World Health Organization (2025). Malaria. [link]. Accessed: 2025-12-02.

Yang, F., Quizon, N., Yu, H., Silamut, K., Maude, R. J., Jaeger, S., and Antani, S. (2020). Cascading YOLO: automated malaria parasite detection for plasmodium vivax in thin blood smears. In Medical Imaging 2020: Computer-Aided Diagnosis, volume 11314, pages 404–410. SPIE.