用傅里叶变换+残差网络识别生成图像的频域指纹，准确率达92.8%。
Fourier-Based GAN Fingerprint Detection using ResNet50
- 将图像转至频域，捕捉生成图像的周期性痕迹。
- 在真实与生成图像上达到92.8%准确率和0.95 AUC。
- 适合数字取证、AI可信系统等工业场景使用。
生成对抗网络（GAN）产生的逼真图像对图像取证和需要内容真实性的工业系统构成严峻挑战。本文结合频域分析与深度学习，解决区分StyleGAN生成图像与真实图像的问题。具体而言，采用二维离散傅里叶变换（2D DFT）将图像转换至傅里叶域，使细微的周期性伪影得以显现。利用ResNet50神经网络对这些变换后的图像进行训练，以区分真实与合成图像。实验表明，该频域模型在准确率上达到92.8%，AUC为0.95，显著优于在原始空间域图像上训练的同类模型。结果表明，生成图像具有独特的频域特征或‘指纹’。该方法凸显了信号处理与深度学习结合在增强数字取证与提升工业AI系统可信度方面的应用潜力。
原文摘要 · Abstract (English)
The rapid rise of photorealistic images produced from Generative Adversarial Networks (GANs) poses a serious challenge for image forensics and industrial systems requiring reliable content authenticity. This paper uses frequency-domain analysis combined with deep learning to solve the problem of distinguishing StyleGAN-generated images from real ones. Specifically, a two-dimensional Discrete Fourier Transform (2D DFT) was applied to transform images into the Fourier domain, where subtle periodic artifacts become detectable. A ResNet50 neural network is trained on these transformed images to differentiate between real and synthetic ones. The experiments demonstrate that the frequency-domain model achieves a 92.8 percent and an AUC of 0.95, significantly outperforming the equivalent model trained on raw spatial-domain images. These results indicate that the GAN-generated images have unique frequency-domain signatures or "fingerprints". The method proposed highlights the industrial potential of combining signal processing techniques and deep learning to enhance digital forensics and strengthen the trustworthiness of industrial AI systems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。