Skip to content
Open access

Deep learning–assisted malaria microscopy with sensitivity-aware threshold optimization

Aug 2026 · Frontiers in Medicine · Vol 13 · 0 citations · 47 references
Medicine

TL;DR

The main finding is that the highest-accuracy classifier is not necessarily the most appropriate sensitivity-oriented screening configuration.

Abstract

Malaria microscopy becomes clinically risky when parasitized erythrocytes are missed, even when a classifier reports high overall accuracy. This study presents an artificial-intelligence-assisted malaria microscopy framework that explicitly optimizes the diagnostic operating point for sensitivity-aware screening. A balanced set of 27,558 National Institutes of Health/National Library of Medicine (NIH/NLM) thin-smear cell images was divided by stratified image-level sampling in the ratio 70:15:15 for training, validation, and independent testing. Two custom convolutional neural networks (CNNs) with complementary capacity–regularization profiles and an equal-weight score-level ensemble were evaluated. Training-only online augmentation, fixed input resizing, validation-only threshold selection, confidence intervals, and error analysis were incorporated to improve reproducibility. At the conventional threshold, the compact CNN achieved the highest accuracy (95.26%). After optimizing the operating point with the recall-weighted F2 objective, the deeper CNN provided the strongest screening-oriented result on the held-out test set, reaching 97.05% sensitivity (95% confidence interval: 96.23%–97.70%), a 96.01% F2-score, a 2.95% false-negative rate, and a receiver-operating-characteristic area under the curve of 0.9876 (95% confidence interval: 0.9842–0.9910). It reduced missed parasitized cells from 177 to 61 relative to its default threshold. The public release does not provide patient identifiers; therefore, the reported split is image-level rather than confirmed to be patient-level independent, and external multicenter validation remains necessary before clinical deployment. The main finding is that the highest-accuracy classifier is not necessarily the most appropriate sensitivity-oriented screening configuration.

Read PDF

Similar papers

Preprint Aug 2026

EMFE: A lightweight, explainable machine learning framework for malaria cell classification

EMFE (Efficient Mathematical Feature Extraction), a five-feature framework for classifying single red-blood-cell images as parasitized or uninfected using Gray World color normalization, adaptive green-channel thresholding, morphological spot detection, and classical machine learning, is presented.

M. A. Al Kafi, Walayat Hussain, Mousumi Karmakar et al. · 0 citations
Review Open access Oct 2026

Deep learning applications in parasite microscopy: A scoping review

Deep learning demonstrates substantial potential for automated parasite detection in microscopy, but key challenges include dataset heterogeneity, lack of standardised evaluation and limited real-world validation.

Rose Mary Maniraj, E. Osman, S. N. H. Sheikh Abdullah et al. · 0 citations
Open access Sep 2026

Rapid multi-species malaria parasite detection using deep learning

Visual examination of Giemsa-stained blood smears remains the gold standard for malaria parasite detection but is labour-intensive, difficult to standardise, and a major barrier to scalable digital microscopy workflows. Towards automating smear counting and supporting digital archiving, we previously developed Plas...

Frank Weate, Yun-Chuan Li, David Novotný et al. · 0 citations
Open access Sep 2026

A Resnet50-Based Machine Learning Diagnostic Model for Malaria Detection and Diagnosis

Malaria remains a major global health challenge, particularly in sub-Saharan Africa, where it accounts for the majority of morbidity and mortality cases. Conventional diagnostic methods, such as microscopic examination of Giemsa-stained blood smears, though considered the gold standard, are time-consuming and prone...

G. Wajiga · 0 citations
Sep 2026

ParasiteNet: Deep Learning-Based Identification of Parasitic Helminth Eggs in Microscopic Images.

The proposed framework demonstrates promising performance for automated parasite egg classification under controlled experimental conditions, however, additional validation using independent external datasets, diverse microscopy settings, and prospective clinical studies is necessary to establish the robustness and gen...

Dolgorsuren Batjargal, Tuvshinsaikhan Tegshee, Temuulen Dorjsuren et al. · 0 citations
Open access Sep 2026

Automated Detection of Malaria Parasites Using Image Processing and Convolutional Neural Network Classifier.

Malaria, being one of the most dangerous parasite diseases endangering human life and leads to high mortality and morbidity rates, affects millions of people, especially in subtropical and tropical regions. Because of the dependence on human skills and the inaccuracy of manual analysis, conventional diagnostic techniqu...

Jamal M. Alrikabi · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.