MosQNet-SA

Explainable Convolutional-Attention Network for Mosquito Classification

MosQNet-SA is a deep learning model for accurate classification of mosquito species (Anopheles, Aedes, Culex), built with a convolutional-attention architecture and explainability via Grad-CAM visualizations.

Key Contributions

  • Convolutional-attention architecture achieving >97% classification accuracy on three mosquito species
  • Explainable predictions with Grad-CAM saliency maps for entomological interpretation
  • RESTful API deployment for real-time dengue and malaria risk mapping in endemic regions
  • Published in PLOS ONE (2026)

References