Watch Your Up-Convolution: CNN Based Generative Deep Neural Networks Are Failing to Reproduce Spectral Distributions

Abstract

Generative convolutional deep neural networks, e.g. popular GAN architectures, are relying on convolution based up-sampling methods to produce non-scalar outputs like images or video sequences. In this paper, we show that common up-sampling methods, i.e. known as upconvolution or transposed convolution, are causing the inability of such models to reproduce spectral distributions of natural training data correctly. This effect is independent of the underlying architecture and we show that it can be used to easily detect generated data like deepfakes with up to 100% accuracy on public benchmarks. To overcome this drawback of current generative models, we propose to add a novel spectral regularization term to the training optimization objective. We show that this approach not only allows to train spectral consistent GANs that are avoiding high frequency errors. Also, we show that a correct approximation of the frequency spectrum has positive effects on the training stability and output quality of generative networks.

Exposing GAN-Generated Faces Using Inconsistent Corneal Specular Highlights

“…As color difference can be easily fixed, more sophisticated detection methods, e.g, [10,16], seek more abstract signal-level traces or fingerprints in the noise residuals to differentiate GAN-synthesized faces. More recent works such as [17,18,19] extend the analysis to the frequency domain, where the upsampling step in the GAN generation leaves specific artifacts. The second category of GAN synthesized face detection methods are of data-driven nature [20,21,22,23,24], where a deep neural network model is trained and employed to classify real and GAN-synthesized faces.…”

Beyond the Spectrum: Detecting Deepfakes via Re-Synthesis

“…GANs have been significantly improved in recent years and are able to synthesize high fidelity images fooling human eyes. However, there persist some problems in GANs revealing the differences between generated distributions and real ones because of commonly-used up-convolution (or deconvolution) operation [Durall et al, 2020], which maps lowresolution tensors to high-resolution ones. Yet these problems are never long-lasting compared to the steady improvement of GANs.…”

“…A CNN classifier is typically trained on the extracted features to perform binary classification for fake detection. However, recent work has shown that such low-level features are rather easy to recognize [Wang et al, 2020] and can be effectively concealed [Durall et al, 2020;Jung and Keuper, 2021]. Due to the steady improvement of generative models and the constantly narrowing gap between real and fake images, this appears not to yield in a reliable and sustainable approach to distinguish real and fake images.…”

DeepFake Videos Detection Using Self-Supervised Decoupling Network

“…Dataset. FaceForensics++ (FF++) [6] is a recently released benchmark dataset and widely used for performance evaluation of face forgery detection [12,21,22]. The dataset consists of 1,000 original videos and their manipulated counterparts created by four typical manipulation methods: Deep-Fakes (DF), Face2Face (F2F), FaceSwap (FS), NeuralTexures (NT).…”