TITLE:
Automated Malaria Detection and Parasitemia Estimation Using an OpenFlexure Microscope with Integrated Autofocus, Slide Scanning, and Deep Learning
AUTHORS:
Daniel Maitethia Memeu, Ezekiel Otieno, Victoria Muthee, Eugene Macharia, Patrick Kubai, Cynthia N. Mugo Mwenda, Dickson Mwenda Kinyua
KEYWORDS:
Parasitemia, Plasmodium, Edge Devices, Point of Care, YOLO, Autofocusing, Slide Scanning
JOURNAL NAME:
Open Journal of Biophysics,
Vol.16 No.3,
July
30,
2026
ABSTRACT: We present a customized OpenFlexure Microscope (OFM) platform for automated malaria detection and parasitemia estimation. The system integrates a Laplacian variance-based autofocus algorithm optimized for high-magnification (100×/1.25 NA oil immersion) imaging, automated slide scanning, and a YOLOv11n deep-learning model for detection of Plasmodium-infected red blood cells (iRBCs), non-infected red blood cells (RBCs), and white blood cells (WBCs). Approximately 4,000 annotated images acquired from Giemsa-stained Plasmodium falciparum thin blood smears were used for model training and evaluation. The proposed autofocus algorithm enabled reliable image acquisition across 100 fields of view, overcoming focus drift observed with the default OpenFlexure autofocus routine. The YOLOv11n model achieved a precision of 70.8%, recall of 91.6%, and F1 score of 79.9% for infected RBC detection. Integrated with automated slide scanning and image analysis, the OFM platform achieved a parasitemia estimation accuracy of approximately 80%, compared with 38% for experienced human microscopists, while reducing analysis time from over 100 minutes to under 40 minutes per slide. The results highlight the challenges associated with reproducible quantitative parasitemia estimation using manual microscopy and demonstrate the advantages of automated image analysis for large-scale cell counting tasks. These findings demonstrate the potential of combining low-cost open-source microscopy with artificial intelligence to provide accurate, reproducible, and scalable malaria parasite detection and parasitemia estimation in resource-limited and point-of-care settings.