CorrGAN: Input Transformation Technique Against Natural Corruptions
Abstract
Because of the increasing accuracy of Deep Neural Networks (DNNs) on different tasks, a lot of real times systems are utilizing DNNs. These DNNs are vulnerable to adversarial perturbations and corruptions. Specifically, natural corruptions like fog, blur, contrast etc can affect the prediction of DNN in an autonomous vehicle. In real time, these corruptions are needed to be detected and also the corrupted inputs are needed to be denoised to be predicted correctly. In this work, we propose CorrGAN approach, which can generate benign input when a corrupted input is provided. In this framework, we train Generative Adversarial Network (GAN) with novel intermediate output-based loss function. The GAN can denoise the corrupted input and generate benign input. Through experimentation, we show that up to 75.2% of the corrupted misclassified inputs can be classified correctly by DNN using CorrGAN.
Cite
Text
Haque et al. "CorrGAN: Input Transformation Technique Against Natural Corruptions." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022. doi:10.1109/CVPRW56347.2022.00032Markdown
[Haque et al. "CorrGAN: Input Transformation Technique Against Natural Corruptions." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022.](https://mlanthology.org/cvprw/2022/haque2022cvprw-corrgan/) doi:10.1109/CVPRW56347.2022.00032BibTeX
@inproceedings{haque2022cvprw-corrgan,
title = {{CorrGAN: Input Transformation Technique Against Natural Corruptions}},
author = {Haque, Mirazul and Budnik, Christof J. and Yang, Wei},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
year = {2022},
pages = {193-196},
doi = {10.1109/CVPRW56347.2022.00032},
url = {https://mlanthology.org/cvprw/2022/haque2022cvprw-corrgan/}
}