{A review of computational techniques for counting Aedes aegypti eggs in ovitrap images

Abstract

Dengue is an infectious viral disease, transmitted by the Aedes mosquito (albopictus and aegypti), which is present in more than 100 countries. Despite the frequent and growing number of confirmed cases, there is no effective vaccine against dengue. The most efficient strategy is monitoring and controlling the vector, performed mainly by cleaning containers that are possible breeding sites. One of the alternatives for monitoring is the estimation of the number of mosquitoes per region. This can be done by counting eggs in images of ovitraps, which are traps composed of a container partially occupied by water or some infusion in which a pallet or blade is inserted, providing a suitable environment for oviposition. The pallet is photographed and the eggs are counted manually. The images contain a variable amount of eggs that can reach thousands of units, making counting work expensive and passive to errors due to human fatigue. An alternative to assist in estimating the number of eggs present in ovitraps are automatic counting methods. The current work presents a bibliographic survey of automatic egg counting works present in the literature, identifying how different techniques are used in solving problems, as well as the factors of greatest influence in the process and possible challenges that still demand solutions.