Application of convolutional neural networks for automated classification of life cycle stages of Plasmodium falciparum in blood smear microscopy

Background. Malaria is a global public health problem. Traditional blood microscopy, the gold standard of diagnosis, is associated with high labor burden, personnel shortages in endemic regions, and subjective assessment. Automation of not only detection but also determination of the pathogen’s developmental stage is critically important for prescribing stage-specific therapy.

Aim. To develop and evaluate the effectiveness of a deep computer vision model for classifying four key stages of the erythrocytic cycle of Plasmodium falciparum based on microphotographs.

Materials and methods. A convolutional neural network with the ResNet-18 architecture was used as the basis. Training was conducted using transfer learning on the PfMorphoBank database, assembled by the authors. The images were divided into four target classes: young trophozoites, mature trophozoites, schizonts, and gametocytes. Classification quality was assessed on validation and held-out test sets using accuracy, macro-precision, macro-recall, and macro-F1.

Results. On the held-out test set, the fine-tuned ResNet-18 achieved 80.83% accuracy; macro-precision was 0.812, macro-recall was 0.808, and macro-F1 was 0.806. Error analysis showed better recognition of mature trophozoites and the greatest difficulty with gametocytes and morphologically similar late-stage forms.

Conclusion. The use of deep neural networks demonstrates high potential in solving the problem of differential diagnosis of malaria parasite stages. The model showed acceptable quality metrics, proving the fundamental feasibility of creating automated prompting systems.

1. Liu Q, Jing WZh, Kang LYu, Liu J, Liu M. Trends of the global, regional and national incidence of malaria in 204 countries from 1990 to 2019 and implications for malaria prevention. Journal of Travel Medicine. 2021; 28(5): taab046. doi: 10.1093/jtm/taab046

2. Sergiev VP, Baranova AM, Kozhevnikova GM, Tokmalayev AK, Chernyshov DV, Chentsov VB, et al. Problems of clinical diagnosis and treatment of P. falciparum malaria in Russian Federation. Terapevticheskii arkhiv. 2018; 90(11): 4–8. doi: 10.26442/terarkh201890114-8

3. Haldar K, Mohandas N. Malaria, erythrocytic infection, and anemia. Hematology. American Society of Hematology. Education Program. 2009: 87–93. doi: 10.1182/asheducation-2009.1.87

4. Shahin OR, Alshammari HH, Alabdali RN, et al. Automated multi-model framework for malaria detection using deep learning and feature fusion. Sci Rep. 2025; 15, 25672. doi: 10.1038/s41598-025-04784-w

5. Fu X, Lin R, Du W, et al. Explainable hybrid transformer for multi-classification of lung disease using chest X-rays. Sci Rep. 2025; 15(1): 6650. doi: 10.1038/s41598-025- 90607-x

6. Saidani O, Umer M, Alturki N, et al. White blood cells classification using multi-fold pre-processing and optimized CNN model. Sci Rep. 2024; 14: 3570. doi: 10.1038/s41598-024-52880-0

7. Mujahid M, Rustam F, Shafique R, et al. Efficient deep learning-based approach for malaria detection using red blood cell smears. Sci Rep. 2024; 14(1): 13249. doi: 10.1038/s41598-024-63831-0

8. Li S, Du Z, Meng X, Zhang Y. Multi-stage malaria parasite recognition by deep learning. Giga Science. 2021; 10(6): giab040. doi: 10.1093/gigascience/giab040

9. Kumar A, Raghunathan A, Jones R, Ma T, Liang P. Fine-tuning can distort pretrained features and underperform out-of-distribution. International Conference on Learning Representations (ICLR). 2022. doi: 10.48550/arXiv.2202.10054

10. Tan D, Liang X. Multiclass malaria parasite recognition based on transformer models and a generative adversarial network. Sci Rep. 2023; 13(1): 17136. doi: 10.1038/s41598-023-44297-y

11. Tekle E, Dese K, Girma S, et al. DeepLeish: a deep learning based support system for the detection of Leishmaniasis parasite from Giemsa-stained microscope images. BMC Med Imaging. 2024; 24(1): 152. doi: 10.1186/s12880-024-01333-1

12. Mura DA, Zedda L, Loddo A, Di Ruberto C. YOLO-Tryppa: A Novel YOLO-Based Approach for Rapid and Accurate Detection of Small Trypanosoma Parasites. J. Imaging. 2025; 11(4): 117. doi: 10.3390/jimaging11040117

13. Lin L, Dacal E, Díez N, et al. Edge Artificial Intelligence (AI) for real-time automatic quantification of filariasis in mobile microscopy. PLoS Negl Trop Dis. 2024; 18(4): e0012117. doi: 10.1371/journal.pntd.0012117