Leveraging Domain Features for Detecting Adversarial Attacks Against Deep Speech Recognition in Noise | IEEE Journals & Magazine | IEEE Xplore