Quantification and Classification Mosquitoes from Images

Vector-borne diseases transmitted by mosquitoes have been neglected over the past years despite constantly affecting the human population around the globe. Female mosquitoes engage in hematophagy to complete their reproductive cycle, but during blood feeding, they can transmit arboviruses such as dengue, chikungunya, and Zika. Controlling the mosquito population is necessary to contain emerging arbovirus epidemics. However, control as an isolated action is not sufficient. Monitoring vector mosquitoes is fundamental in combating arboviruses, which involves estimating the mosquito population in a region. Despite various monitoring techniques, a significant limiting factor remains the need for a specialized professional (entomologist) to identify and count the captured mosquito species, making the monitoring process slow, imprecise, and manual. Thus, the field of artificial intelligence and machine learning, with tasks of quantification and classification, can assist in automating the mosquito monitoring process. In this project, we explore machine-learning techniques and low-cost hardware to build mechanisms for detecting invasive species and counting disease vector.