Adversarial Examples for Edge Detection: They Exist, and They Transfer
This reveals a security vulnerability in edge detection and related vision models, with potential implications for safety-critical applications, though it is incremental as it extends known adversarial example issues to a new domain.
The paper demonstrates that convolutional neural network-based edge detectors, such as HED, are vulnerable to adversarial examples, where small image perturbations cause failures like missing edges, false detections, or hallucinated edges, and these attacks transfer to other CNN models, reducing accuracy in tasks like image classification and semantic segmentation.
Convolutional neural networks have recently advanced the state of the art in many tasks including edge and object boundary detection. However, in this paper, we demonstrate that these edge detectors inherit a troubling property of neural networks: they can be fooled by adversarial examples. We show that adding small perturbations to an image causes HED, a CNN-based edge detection model, to fail to locate edges, to detect nonexistent edges, and even to hallucinate arbitrary configurations of edges. More surprisingly, we find that these adversarial examples transfer to other CNN-based vision models. In particular, attacks on edge detection result in significant drops in accuracy in models trained to perform unrelated, high-level tasks like image classification and semantic segmentation. Our code will be made public.