In this article, we explore the overview of epistemic uncertainty in deep learning classifiers and develop an intuitive approach to detecting the presence of specific image instances using an ensemble of models.