Answers are grounded in this paper and cite their sources.
Mosquito-borne diseases (MBDs) remain a major global health challenge. Vector surveillance, pathogen detection, and clinical diagnosis are central to control efforts but are limited by labor-intensive workflows and shortages of trained personnel. Computer vision (CV) offers opportunities for scalable automation throughout the MBD management pipeline. We aimed to systematically characterize the applications of computer vision in mosquito-borne diseases research, summarize methodological developments, and identify opportunities and barriers to translating these advances into public health practice. We conducted a systematic review in accordance with PRISMA 2020 guidelines, searching PubMed, Scopus, and IEEE Xplore for studies published between January 1, 2010, and December 31, 2025, that applied computer vision to any aspect of MBD research or control. A total of 1,521 records were screened, and 650 studies met the inclusion criteria. Studies were categorized by computer vision application area, data types, algorithmic approaches, and data acquisition methods. Descriptive quantitative analysis and qualitative thematic synthesis were performed. The use of computer vision in MBD research increased significantly over the review period. Most studies focused on diagnostic applications, particularly Plasmodium detection in malaria microscopy. More recent work incorporated deep learning models for mosquito species classification, UAV- and satellite-based habitat mapping, and smartphone-based non-invasive diagnostic imaging. Emerging areas included 3D mosquito behavioral analysis and Internet of Things-enabled surveillance. Algorithmic approaches evolved from traditional image processing to convolutional neural networks, You Only Look Once (YOLO)-based object detectors, and transformer architectures. However, most models were developed using small, geographically-limited datasets, lacked external or prospective validation, and were rarely evaluated in operational settings, limiting their translational readiness.
No takes yet. Share an insight, caveat, or question.
Ong et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: