GAP: Quantifying the Generative Adversarial Set and Class Feature Applicability of Deep Neural Networks | IEEE Conference Publication | IEEE Xplore