Detecção de Potenciais Focos do Aedes aegypti em Vídeos Aéreos Usando Redes Neurais

The Aedes aegypti mosquito is the transmitter of diseases like dengue, zika, chikungunya, and urban yellow fever, which affect millions of people every year. The main control for these diseases is done by searching and eliminating the mosquitos breeding grounds. In this work we introduce a comprehensive database of aerial videos recorded with a drone, where all objects of interest are identified by their respective bounding boxes, and describe an object detection system based on deep neural networks. We employ phase correlation to obtain the spatial alignment between the video frames and track the objects. By doing so, we are capable of registering the detected objects, minimizing false positives and correcting most false negatives. Using the ResNet-101-C4 as backbone, it is possible to obtain 0.74 in terms of F1-score on the proposed dataset.