Mosquitoes Classification using EfficientNetB4 Transfer Learning Model

Abstract

A million individuals per year die from diseases spread by mosquitoes. When a mosquito stings, saliva is injected into the body, causing the illness to spread to the victims. In a surveillance programmed for infections propagated through mosquito detection, categorization is the most crucial stage. Classification and labeling are difficult and time-consuming procedures when employing traditional method to collect data. Transfer learning is an advanced image processing techniques that offers a great solution to this problem. With very few training images, transfer learning is a form of CNN that can be beneficial and long-lasting for image analysis. This research will enhance human health and quality of life. The purpose of this approach is to create a systematic process for developing a categorization system using an EfficentNetB4 transfer learning algorithm for mosquitoes. The resultant performance analysis showed that the EfficentNetB4 model offers an accuracy of 85.79%, loss of 40.05%, val_loss of 40.42%, and val_accuracy of 86.30%.