Guarding Against Adversarial Attacks using Biologically Inspired Contour Integration
Salman Hameed Khan, Alexander K.C. Wong, Bryan Tripp · Journal of Computational Vision and Imaging Systems · 2018
Artificial vision systems are susceptible to adversarial attacks. Smallintentional changes to images can cause these systems to mis-classify with high confidence. The brain has many mechanisms forstrengthening weak or confusing inputs. One such technique, con-tour integration can separate objects from irrelevant background.We show that incorporating contour integration within artificial vi-sual systems can increase their robustness to adversarial attacks.