lihaod / GAN_image_detection

Implementation of "Detection of Deep Network Generated Images Using Disparities in Color Components"

Identification of GAN Generated Images

This is an implementation of the paper Identification of deep network generated images using disparities in color components.

Requirements

• Matlab R2015b+

Usage

Simply call the function gan_img_detection_fea to calculate the features proposed in the paper.

Fea = gan_img_detection_fea(Img)

Note: To train and test on a dataset, please use the code of ensembles (for sample-aware and model-aware detection) / SVM (for model-unaware detection) classifiers.

Citation

If you use our code please cite:

@article{li2020identification,
  title={Identification of Deep Network Generated Images Using Disparities in Color Components},
  author={Li, Haodong and Li, Bin and Tan, Shunquan and Huang, Jiwu},
  journal={Signal Processing},
  volume = {174},
  pages={107616},
  year={2020},
  issn = {0165-1684},
  doi = {https://doi.org/10.1016/j.sigpro.2020.107616},
}