Data to 7 Oct 2026

Intriguing properties of synthetic images: from generative adversarial networks to diffusion models

arXiv 2304.06408 ↗First posted 13 Apr 2023cs.CVcited by 121

On the map: Diffusion models → Image generation, text-to-image and editing

Abstract (arXiv)

Detecting fake images is becoming a major goal of computer vision. This need is becoming more and more pressing with the continuous improvement of synthesis methods based on Generative Adversarial Networks (GAN), and even more with the appearance of powerful methods based on Diffusion Models (DM). Towards this end, it is important to gain insight into which image features better discriminate fake images from real ones. In this paper we report on our systematic study of a large number of image generators of different families, aimed at discovering the most forensically relevant characteristics of real and generated images. Our experiments provide a number of interesting observations and shed light on some intriguing properties of synthetic images: (1) not only the GAN models but also the DM and VQ-GAN (Vector Quantized Generative Adversarial Networks) models give rise to visible artifacts in the Fourier domain and exhibit anomalous regular patterns in the autocorrelation; (2) when the dataset used to train the model lacks sufficient variety, its biases can be transferred to the generated images; (3) synthetic and real images exhibit significant differences in the mid-high frequency signal content, observable in their radial and angular spectral power distributions.

Key facts. Cited by 121 papers in the map (data through 7 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 43 of its 66 references that are in the map. First posted on arXiv 13 Apr 2023.

Builds on

43 of 66 references.

Closest ancestors · 6

Same direction, newest first

Other references · 37

2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
2020Denoising Diffusion Probabilistic Modelscited by 20,851
2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
2022Scalable Diffusion Models with Transformerscited by 5,931
2017Neural Discrete Representation Learningcited by 4,650
2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
2020Taming Transformers for High-Resolution Image Synthesiscited by 3,229
2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 2,994

25 more

2017Progressive Growing of GANs for Improved Quality, Stability, and Variationcited by 2,967
2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
2021LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairscited by 1,544
2016Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoisingcited by 1,352
2021Alias-Free Generative Adversarial Networkscited by 1,181
2020Leveraging Frequency Analysis for Deep Fake Image Recognitioncited by 501
2023Scaling up GANs for Text-to-Image Synthesiscited by 492
2023Towards Universal Fake Image Detectors that Generalize Across Generative Modelscited by 483
2019Detecting and Simulating Artifacts in GAN Fake Imagescited by 293
2020Watch your Up-Convolution: CNN Based Generative Deep Neural Networks are Failing to Reproduce Spectral Distributionscited by 278
2023StyleGAN-T: Unlocking the Power of GANs for Fast Large-Scale Text-to-Image Synthesiscited by 212
2020Focal Frequency Loss for Image Reconstruction and Synthesiscited by 204
2020Artificial Fingerprinting for Generative Models: Rooting Deepfake Attribution in Training Datacited by 200
2018Do GANs leave artificial fingerprints?cited by 188
2021Are GAN generated images easy to detect? A critical analysis of the state-of-the-artcited by 139
2019Fourier Spectrum Discrepancies in Deep Network Generated Imagescited by 138
2022Diffusion Probabilistic Model Made Slimcited by 121
2018Noiseprint: a CNN-based camera model fingerprintcited by 114
2021A Closer Look at Fourier Spectrum Discrepancies for CNN-generated Images Detectioncited by 68
2021Reverse Engineering of Generative Models: Inferring Model Hyperparameters from Generated Imagescited by 58
2021Eyes Tell All: Irregular Pupil Shapes Reveal GAN-generated Facescited by 48
2022Lighting (In)consistency of Paint by Textcited by 35
2022Perspective (In)consistency of Paint by Textcited by 34
2021More Real than Real: A Study on Human Visual Perception of Synthetic Facescited by 23
2022From Points to Functions: Infinite-dimensional Representations in Diffusion Modelscited by 11

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 7; 2024: 40; 2025: 45; 2026: 29.

121 papers.

Sort:

By area:

2023Raising the Bar of AI-generated Image Detection with CLIPcited by 163
2024AEROBLADE: Training-Free Detection of Latent Diffusion Images Using Autoencoder Reconstruction Errorcited by 123
2024Leveraging Representations from Intermediate Encoder-blocks for Synthetic Image Detectioncited by 88
2024Detecting Multimedia Generated by Large AI Models: A Surveycited by 73
2024A Bias-Free Training Paradigm for More General AI-generated Image Detectioncited by 67
2023Towards Understanding the Interplay of Generative Artificial Intelligence and the Internetcited by 54
2024Zero-Shot Detection of AI-Generated Imagescited by 50
2024Real-Time Deepfake Detection in the Real-Worldcited by 49
2023Generalizable Synthetic Image Detection via Language-guided Contrastive Learningcited by 48
2024LaRE2: Latent Reconstruction Error Based Method for Diffusion-Generated Image Detectioncited by 47
2024Fake or JPEG? Revealing Common Biases in Generated Image Detection Datasetscited by 43
2025CO-SPY: Combining Semantic and Pixel Features to Detect Synthetic Images by AIcited by 42

38 more

2024On Learning Multi-Modal Forgery Representation for Diffusion Generated Video Detectioncited by 40
2025D3: Training-Free AI-Generated Video Detection Using Second-Order Featurescited by 32
2024Contrasting Deepfakes Diffusion via Contrastive Learning and Global-Local Similaritiescited by 27
2024Frequency Masking for Universal Deepfake Detectioncited by 23
2025Seeing What Matters: Generalizable AI-generated Video Detection with Forensic-Oriented Augmentationcited by 23
2024Towards a Universal Synthetic Video Detector: From Face or Background Manipulations to Fully AI-Generated Contentcited by 22
2024Mixture of Low-rank Experts for Transferable AI-Generated Image Detectioncited by 21
2023The Age of Synthetic Realities: Challenges and Opportunitiescited by 15
2024Are CLIP features all you need for Universal Synthetic Image Origin Attribution?cited by 14
2024ForgeLens: Data-Efficient Forgery Focus for Generalizable Forgery Image Detectioncited by 14
2024Synthetic Image Verification in the Era of Generative AI: What Works and What Isn't There Yetcited by 13
2025Characterizing Photorealism and Artifacts in Diffusion Model-Generated Imagescited by 13
2024A Survey of Defenses Against AI-Generated Visual Media: Detection,Disruption, and Authenticationcited by 12
2024Vulnerabilities in AI-generated Image Detection: The Challenge of Adversarial Attackscited by 12
2024StealthDiffusion: Towards Evading Diffusion Forensic Detection through Diffusion Modelcited by 12
2025Spectral Analysis of Diffusion Models with Application to Schedule Designcited by 12
2025Survey on AI-Generated Media Detection: From Non-MLLM to MLLMcited by 12
2023M3Dsynth: A dataset of medical 3D images with AI-generated local manipulationscited by 11
2024Model Collapse in the Self-Consuming Chain of Diffusion Finetuning: A Novel Perspective from Quantitative Trait Modelingcited by 11
2024Exploring the Adversarial Robustness of CLIP for AI-generated Image Detectioncited by 11
2024SGEdit: Bridging LLM with Text2Image Generative Model for Scene Graph-based Image Editingcited by 10
2025AI-GenBench: A New Ongoing Benchmark for AI-Generated Image Detectioncited by 10
2025LOTA: Bit-Planes Guided AI-Generated Image Detectioncited by 10
2025WILD: a new in-the-Wild Image Linkage Dataset for synthetic image attributioncited by 9
2025Navigating the Challenges of AI-Generated Image Detection in the Wild: What Truly Matters?cited by 8
2024Detecting Origin Attribution for Text-to-Image Diffusion Modelscited by 7
2025Explainable Synthetic Image Detection through Diffusion Timestep Ensemblingcited by 7
2025SelfMAD: Enhancing Generalization and Robustness in Morphing Attack Detection via Self-Supervised Learningcited by 7
2025Riemannian-Geometric Fingerprints of Generative Modelscited by 7
2024DeepFeatureX Net: Deep Features eXtractors based Network for discriminating synthetic from real imagescited by 6
2024Evolving from Single-modal to Multi-modal Facial Deepfake Detection: Progress and Challengescited by 6
2024Continuous fake media detection: adapting deepfake detectors to new generative techniquescited by 6
2024Deepfake Media Forensics: State of the Art and Challenges Aheadcited by 6
2026When Pretty Isn't Useful: Investigating Why Modern Text-to-Image Models Fail as Reliable Training Data Generatorscited by 6
2023Model Synthesis for Zero-Shot Model Attributioncited by 5
2023Synthetic Image Detection: Highlights from the IEEE Video and Image Processing Cup 2022 Student Competitioncited by 5
2024Artifact Feature Purification for Cross-domain Detection of AI-generated Imagescited by 5
2024Are Images Indistinguishable to Humans Also Indistinguishable to Classifiers?cited by 5

50 most cited of 121 All 121 (JSON)

Similar, not cited

Close in wording, no citation either way