Speaker
Description
Mosquito-borne diseases remain a pressing global health challenge, particularly in regions where effective surveillance and control of vector populations are limited. Accurate and timely identification of mosquito genera is essential for designing targeted interventions, as species such as Aedes, Anopheles, and Culex are responsible for transmitting life-threatening diseases including malaria, dengue, yellow fever, and Zika. Early and precise detection of these vectors is critical for breaking transmission cycles, guiding vector control programs, and reducing disease burden in vulnerable communities. This study presents a deep learning-based object detection framework employing YOLOv5 to detect and classify the three major mosquito genera alongside non-mosquito classes. An image dataset was developed from open-source repositories and augmented to improve model robustness. The trained YOLOv5s model achieved excellent performance, with mAP@0.5 of 0.993 and mAP@0.5:0.95 of 0.654, demonstrating both high detection accuracy and strong generalization across IoU thresholds. The model maintained a detection confidence range of 66% to 92% and proved lightweight, with 7 million parameters and 15.8 GFLOPs. By enabling reliable real-time mosquito identification, this approach supports improved public health decision-making, enhances early warning systems, and provides a cost-effective solution for vector surveillance in resource-limited settings. These results highlight the model’s potential for deployment on mobile and IoT devices to strengthen vector monitoring and contribute to global efforts in reducing mosquito-borne disease transmission.