Journal
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022)
Volume -, Issue -, Pages 18941-18952Publisher
IEEE COMPUTER SOC
DOI: 10.1109/CVPR52688.2022.01839
Keywords
-
Funding
- EPFL EssentialTech Centre Humanitarian Action Challenge Grant
- ETH4D
Ask authors/readers for more resources
We introduce a set of image transformations that take into account the geometry of the scene, making them more realistic compared to existing approaches. These transformations can be used to evaluate the robustness of models and serve as data augmentation mechanisms for training neural networks. The study shows that incorporating these transformations can expose vulnerabilities in existing models and improve the robustness of models when used as 3D data augmentation.
We introduce a set of image transformations that can be used as corruptions to evaluate the robustness of models as well as data augmentation mechanisms for training neural networks. The primary distinction of the proposed transformations is that, unlike existing approaches such as Common Corruptions [27], the geometry of the scene is incorporated in the transformations - thus leading to corruptions that are more likely to occur in the real world. We also introduce a set of semantic corruptions (e.g. natural object occlusions. See Fig. 1). We show these transformations are 'efficient' (can be computed on-the-fly), 'extendable' (can be applied on most image datasets), expose vulnerability of existing models, and can effectively make models more robust when employed as '3D data augmentation' mechanisms. The evaluations on several tasks and datasets suggest incorporating 3D information into benchmarking and training opens up a promising direction for robustness research.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available