Student Research Reports

View Research Report

Mosquito-borne diseases (MBDs), such as dengue virus, chikungunya virus, and West Nile virus, cause over one million deaths globally every year. Because many such diseases are spread by the Aedes and Culex mosquitoes, tracking these larvae is critical to mitigating the spread of MBDs. Even as citizen science projects to obtain large mosquito image datasets continuously grow, the manual annotation of mosquito images is becoming ever more time-consuming and inefficient. Previous research has seen computer vision used to identify mosquito species, and the Convolutional Neural Network (CNN) has become the de-facto for image classification. But, these models typically require substantial computational resources. This research introduces the application of the Vision Transformer (ViT) in a comparative study to improve image classification on Aedes and Culex larvae. By using mosquito larvae image data from the GLOBE Observer Mosquito Habitat Mapper, two ViT models, ViT-Base and CvT-13, and two CNN models, ResNet-18 and ConvNeXT, from the HuggingFace library were trained and compared to determine the most effective model to classify mosquito larvae as Aedes, Culex, or neither. Testing revealed that ConvNeXT obtained the greatest values across all four classification metrics, making it a viable method for mosquito larvae image classification. Based on the results of this work, future research could include creating and implementing a model specifically designed for mosquito larvae classification by combining elements of CNN and transformer architecture. Keywords: Vision Transformer, Convolutional Neural Network, Image Classification, Mosquito-Borne Disease.