Desenvolvimento de api para aprendizado profundo de imagens de pequenos ruminantes

Abstract

Sheep and goat farming has always been very important to Brazil’s northeastern region, being a key part for the local economy. These livestocks, when raised in confinement, can present many sicknesses and parasitosis, like Haemonchus contortus, that may cause anemia and important losses to the local farmers. There are methods to control anemia in sheep and goat livestocks, one of them is FAMACHA© method, which aims to clinically identify resistant animals, resilient and sensitive to parasitic infections, optimizing the treatment selectively in real situations at the countryside, without the need of using laboratory resources. (MOLENTO et al., 2004). The study of this document aims to create a machine learning model which classifies animal’s mucosa as anemic or non-anemic using convolutional neural networks as well as deploying this model in the production environment through Python frameworks.