1. Introduction

INTRODUCTION

I came up with the abstract for this project due to family members who suffered from vector-borne illnesses. Although my mother and grandmother were fortunate enough to afford appropriate healthcare, this was not the case for the majority of the population where they lived. I thought that it might be an interesting subject to pursue through the field of computer vision, and eventually decided to follow through with the project due to the significant impact infectious diseases has had on the world in the past years. I hope that this project will help the general population with disease control, to reduce the chances of an outbreak of a disease carried by insects or other animals. As the pandemic has influenced the general population negatively in multiple areas, such as mortality rates and mental health, all the way to education, I considered it a relevant and potentially useful topic to study. I hoped to use artificial intelligence (more specifically, neural networks) to identify if an insect in a picture was a vector carrier. There has been previous research in the field, however, this has only been used to identify the aedes mosquito. This app is called Mosquito Alert.

Another project was the detection of the aedes mosquito through audio classification. The data used to train the network was the audio of the wingbeats of the aedes mosquito, and used a spectrogram to ‘represent the mosquito wingbeat frequency over time visually’. (Detecting Aedes Aegypti Mosquitoes through Audio Classification with Convolutional Neural Networks).This model achieved an accuracy of 97%.

'The incidence of mosquito-borne diseases is significant in under-developed regions, mostly due to the lack of resources to implement aggressive control measurements against mosquito proliferation. A potential strategy to raise community awareness regarding mosquito proliferation is building a live map of mosquito incidences using smartphone apps and crowdsourcing.' -Marcelo Schreiber Fernandes, Weverton Cordeiro, Mariana Recamonde-Mendoza

PICAT is an integrated platform that controls viruses in Catalonia; they use data such as geolocation data to predict cases using simulations and other visualization tools. I hope that my project will build upon this by adding various different species of mosquitos such as the culex or anopheles, as well as other genera such as ticks or flies.