A systematic framework for natural perturbations from videos

24 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

On Robustness and Transferability of Convolutional Neural Networks (2020, methodology): “Attempting to focus on naturally occurring invariances, [54] annotated two video datasets: IMAGENET-VIDROBUST and YOUTUBE-BB-ROBUST, derived from the IMAGENET-VID [11] and YOUTUBE-BB [51] datasets respectively.”
Self-Supervised Learning of Video-Induced Visual Invariances (2019, methodology): “To measure if this is indeed the case, we evaluate our VIVI-Ex(4)-Co(100%) model on the ImageNet-Vid-Robust [54] benchmark.”
Test-Time Training for Out-of-Distribution Generalization (2019, background): “These include images with diverse types of corruption at various levels (Hendrycks & Dietterich, 2019), video frames of moving objects (Shankar et al., 2019), and a new test set of unknown distribution shifts collected by Recht et al. (2018).”
Delving Deeper into Anti-Aliasing in ConvNets (2020, background): “For example, a tiny shift in the input image can lead to drastic changes in the output prediction of convolutional neural networks (ConvNets) [1, 28, 31].”
Evaluating Machine Accuracy on ImageNet (2020, background): “Other forms of natural distribution shift such as robustness to temporal changes could be one candidate (Gu et al., 2019; Shankar et al., 2019).”