Optimizing Deep Learning for Facemask Detection: A Comparative Study of Optimizers

During the COVID era, a significant influence on our well-being was the choice to don face masks, which played a crucial role in curbing viral transmission and promoting overall hygiene. The current study centers around the utilization of Deep Neural Networks to accomplish the task of detecting facemasks. However, this endeavor demands a considerable investment of time and substantial computational resources for effective training. Our investigation is particularly concerned with the synergy between the chosen optimizer and the pooling layer, aiming to yield a precise outcome that discerns the presence or absence of a facemask on an individual. Among the various configurations, the combination of Adam optimizer and Average pooling layer is the most effective, attaining an impressive accuracy rate of 96.25%.