[RE] CNN-generated images are surprisingly easy to spot...for now

TL;DR

This paper evaluates the reproducibility of methods for detecting CNN-generated images, focusing on data augmentation and diversity to improve generalization across unseen architectures and identifying their limitations.

Contribution

It assesses the effectiveness of data augmentation and dataset diversity in enhancing classifier generalization for CNN-generated image detection.

Findings

Data augmentation and diverse datasets improve detection generalization.

Limitations exist in the current techniques' ability to handle all unseen CNN architectures.

Reproducibility of the original methods is confirmed.

Abstract

This work evaluates the reproducibility of the paper "CNN-generated images are surprisingly easy to spot... for now" by Wang et al. published at CVPR 2020. The paper addresses the challenge of detecting CNN-generated imagery, which has reached the potential to even fool humans. The authors propose two methods which help an image classifier to generalize from being trained on one specific CNN to detecting imagery produced by unseen architectures, training methods, or data sets. The paper proposes two methods to help a classifier generalize: (i) utilizing different kinds of data augmentations and (ii) using a diverse data set. This report focuses on assessing if these techniques indeed help the generalization process. Furthermore, we perform additional experiments to study the limitations of the proposed techniques.