Xavier Boix

Foveation-based Mechanisms Alleviate Adversarial Examples

2015 · arXiv (Cornell University)

We show that adversarial examples, i.e., the visually imperceptible perturbations that result in Convolutional Neural Networks (CNNs) fail, can be alleviated with a mechanism based on foveations---applying the CNN in different image regions. To see …

When and how CNNs generalize to out-of-distribution category-viewpoint combinations

2020 · arXiv (Cornell University)

Object recognition and viewpoint estimation lie at the heart of visual understanding. Recent works suggest that convolutional neural networks (CNNs) fail to generalize to out-of-distribution (OOD) category-viewpoint combinations, ie. combinations not seen during training. In …