ImageNet-C

ImageNet-C is a benchmark dataset that contains corrupted versions of images from the ImageNet dataset and is used to evaluate the robustness of models against various types of image distortions.

Papers using ImageNet-C (3)

VISAT: Benchmarking Adversarial and Distribution Shift Robustness in Traffic Sign Recognition with Visual Attributes2025Medbn: Robust Test-time Adaptation Against Malicious Test Samples2024 · 2 citesBenchmarking Robustness To Adversarial Image Obfuscations2023 · 2 cites