Fooling blind image quality assessment by optimizing a human-understandable color filter

This paper presents the submission of our RU-DS team to the Pixel Privacy Task 2020. We propose to fool the blind image quality assessment model by transforming images based on optimizing a human-understandable color filter. In contrast to the common work that relies on small, L_p-bounded additive pixel perturbations, our approach yields large yet smooth perturbations. Experimental results demonstrate that in the specific context of this task, our approach is able to achieve strong adversarial effects, but has to sacrifice the image appeal.