ImageNet-D: Benchmarking Neural Network Robustness on Diffusion Synthetic Object
This work addresses the need for more rigorous and diverse robustness testing in computer vision, though it is incremental as it builds on existing benchmarks by using generative models.
The authors tackled the problem of evaluating neural network robustness by introducing ImageNet-D, a benchmark using diffusion models to generate diverse synthetic images, which caused accuracy drops of up to 60% in various vision models including ResNet, CLIP, and MiniGPT-4.
We establish rigorous benchmarks for visual perception robustness. Synthetic images such as ImageNet-C, ImageNet-9, and Stylized ImageNet provide specific type of evaluation over synthetic corruptions, backgrounds, and textures, yet those robustness benchmarks are restricted in specified variations and have low synthetic quality. In this work, we introduce generative model as a data source for synthesizing hard images that benchmark deep models' robustness. Leveraging diffusion models, we are able to generate images with more diversified backgrounds, textures, and materials than any prior work, where we term this benchmark as ImageNet-D. Experimental results show that ImageNet-D results in a significant accuracy drop to a range of vision models, from the standard ResNet visual classifier to the latest foundation models like CLIP and MiniGPT-4, significantly reducing their accuracy by up to 60\%. Our work suggests that diffusion models can be an effective source to test vision models. The code and dataset are available at https://github.com/chenshuang-zhang/imagenet_d.