Fetching the paper…

Semantic Adversarial Attacks: Parametric Transformations That Fool Deep Classifiers · Around