Impact of Data Domain on FSR Training

The objective of this repository is to test the transferability of the FSR module proposed in Feature Separation and Recalibration for Adversarial Robustness.

The codes are modified based on the original implementation of FSR.

References

1. Kim, Woo Jae, Yoonki Cho, Junsik Jung, and Sung-Eui Yoon. "Feature Separation and Recalibration for Adversarial Robustness." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023.