Wavelet-based GAN Fingerprint Detection using ResNet50

TL;DR

This paper introduces a wavelet-based method using ResNet50 to effectively detect GAN-generated images by exploiting subtle frequency artifacts, achieving higher accuracy than spatial domain models.

Contribution

The study demonstrates that wavelet preprocessing combined with ResNet50 significantly improves GAN image detection accuracy over traditional spatial domain approaches.

Findings

Wavelet-based models achieved over 93% accuracy.

Daubechies wavelet outperforms Haar in detection.

Wavelet artifacts serve as unique fingerprints for GAN images.

Abstract

Identifying images generated by Generative Adversarial Networks (GANs) has become a significant challenge in digital image forensics. This research presents a wavelet-based detection method that uses discrete wavelet transform (DWT) preprocessing and a ResNet50 classification layer to differentiate the StyleGAN-generated images from real ones. Haar and Daubechies wavelet filters are applied to convert the input images into multi-resolution representations, which will then be fed to a ResNet50 network for classification, capitalizing on subtle artifacts left by the generative process. Moreover, the wavelet-based models are compared to an identical ResNet50 model trained on spatial data. The Haar and Daubechies preprocessed models achieved a greater accuracy of 93.8 percent and 95.1 percent, much higher than the model developed in the spatial domain (accuracy rate of 81.5 percent). The Daubechies-based model outperforms Haar, showing that adding layers of descriptive frequency patterns can lead to even greater distinguishing power. These results indicate that the GAN-generated images have unique wavelet-domain artifacts or "fingerprints." The method proposed illustrates the effectiveness of wavelet-domain analysis to detect GAN images and emphasizes the potential of further developing the capabilities of future deepfake detection systems.