The Importance of Neural Networks in Aiding the Identification of Monkeypo

Resumo

DOI 10.5281/zenodo.10713093

With the arrival of covid-19 and a global quarantine scenario, the use of technology in 
the healthcare sector has grown dramatically, and telemedicine has become a great resource for  healthcare professionals and patients. It is estimated that demand for the service increased by 226% in the first two months of 2022, compared to the previous year. However, after the 
preventive measures and the decrease in cases and deaths, telemedicine has remained a great ally for its many benefits, such as the speed of care, less exposure to contagious diseases, and no travel  costs. COVID-19 was not the only contagious disease during this pandemic period, while the focus  of governments and people was on controlling and reducing covid-19 cases, other diseases were boosting their numbers, such as monkeypox. 
Shortly after the covid crisis, there was a huge increase in monkeypox, a disease caused by 
by the orthopoxvirus genus. It is a disease characterized by skin rashes and can be transmitted by physical contact and airborne droplets, so the patient also needs to be isolated so that there is no contagion. The focus of our study is the comparison between two neural networks, Resnet18 and Xception, in the context of image recognition, to identify symptoms on patients' skin. Notably, Resnet18 has emerged as the option with the best cost-benefit ratio. This choice provides convenience, speed, and safety, ensuring effectiveness in symptom identification without unnecessarily exposing the health condition.