文献汇总｜AI生成图像模型溯源相关工作汇总（2019年至今）

前言：本篇博客总结 2019 年至今通用AIGI（AI-Generated Image）模型溯源相关研究工作。

–
AIGI溯源相关博客如下：

文献研读｜AIGC溯源场景及研究进展

2019

Attributing Fake Images to GANs: Learning and Analyzing GAN Fingerprints. ICCV 2019
Ning Yu, Larry Davis, Mario Fritz. University of Maryland, USA
Do GANs Leave Artificial Fingerprints? MIPR，2019.
Francesco Marra, Diego Gragnaniello, Luisa Verdoliva, Giovanni Poggi. University Federico II of Naples, Italy
核心思想：基于GAN噪声残差相似度的检测。受相机固有缺陷导致的光响应非均匀性（PRNU）启发，提取GAN生成图像中的指纹，借此完成溯源
Incremental learning for the detection and classification of GAN-generated image. WIFS, 2019
Francesco Marra; Cristiano Saltori; Giulia Boato; Luisa Verdoliva. University Federico II of Naples, Italy
Source Generator Attribution via Inversion. CVPR Workshops 2019
Michael Albright and Scott McCloskey. Honeywell ACST
核心思想：给定一个probe image，和多个GANs，通过损失优化的方式，找到最佳结果对应的潜向量，选择重建效果最接近probe image的GAN
Scalable fine-grained generated image classification based on deep metric learning. arXiv, 2019
Xuan, Xinsheng and Peng, Bo and Wang, Wei and Dong, Jing. Chinese Academy of Sciences, China

Attributing and Detecting Fake Images Generated by Known GANs. SPW, 2020
Matthew Joslin and Shuang Hao. University of Texas at Dallas, USA
核心思想：GAN图像FFT特征的平均作为fingerprint，然后和 test image FFT 计算相似度

Towards discovery and attribution of open-world gan generated images. ICCV 2021
Sharath Girish, Saksham Suri, Saketh Rambhatla, Abhinav Shrivastava. University of Maryland, USA
Does a GAN leave distinct model-specific fingerprints?. BMVC 2021
Yuzhen Ding, Nupur Thakur and baoxin Li. Arizona State University, USA
Deepfake attribution: On the source identification of artificially generated images. Wires, 2021
Brandon Khoo, Raphaël C.-W. Phan, Chern-Hong Lim. University, Malaysia, Malaysia
On attribution of deepfakes. arXiv 20210303
Baiwu Zhang, Jin Peng Zhou, Ilia Shumailov, Nicolas Papernot. University of Toronto, Canada
核心思想：提出Strict Attribution 和 Relaxed attribution 场景，想法和Michael Albright的那篇类似

RepMix: Representation Mixing for Robust Attribution of Synthesized Images. ECCV 2022
Tu Bui, Ning Yu, and John Collomosse. University of Surrey, UK
Deepfake Network Architecture Attribution. AAAI 2022
Tianyun Yang, Ziyao Huang, Juan Cao, Lei Li, Xirong Li. Chinese Academy of Sciences, China
核心思想：提出DNA-Det，一阶段使用自监督学习使模型学习用于分类的特征，二阶段实施GAN架构分类。两阶段均使用对比学习
Did you use my GAN to generate fake? post-hoc attribution of GAN images via latent recovery. IJCNN 2022
Syou Hirofumi; Kazuto Fukuchi; Yohei Akimoto; Jun Sakuma. Univresity of Tskuba, Japan

Where Did I Come From? Origin Attribution of AI-Generated Images. NeurIPS, 2023
Zhenting Wang, Chen Chen, Yi Zeng, Lingjuan Lyu, Shiqing Ma. Rutgers University, USA
DE-FAKE- Detection and Attribution of Fake Images Generated by Text-to-Image Generation Models. CCS, 2023
Zeyang Sha, Zheng Li, Ning Yu, Yang Zhang. CISPA, Gernamy
Reverse engineering of generative models: Inferring model hyperparameters from generated images.IEEE TPAMI 2023
Vishal Asnani, Xi Yin, Tal Hassner, Xiaoming Liu. Michigan State University, USA
Open Set Synthetic Image Source Attribution. BMVC 2023
Shengbang Fang, Tai D. Nguyen, Matthew C. Stamm. Drexel University, USA
Progressive Open Space Expansion for Open-Set Model Attribution. CVPR 2023
Tianyun Yang, Danding Wang, Fan Tang, Xinying Zhao, Juan Cao, Sheng Tang. Chinese Academy of Sciences, China
Open Set Classification of GAN-based Image Manipulations via a ViT-based Hybrid Architecture. CVPR Workshops 2023
Jun Wang, Omran Alamayreh Benedetta Tondi Mauro Barni. University of Siena, Italy

Single-Model Attribution of Generative Models Through Final-Layer Inversion. ICML, 2024
Mike Laszkiewicz, Jonas Ricker, Johannes Lederer, Asja Fischer. Ruhr University Bochum, Germany
How to Trace Latent Generative Model Generated Images without Artificial Watermark? ICML, 2024
Zhenting Wang, Vikash Sehwag, Chen Chen, Lingjuan Lyu, Dimitris N. Metaxas, Shiqing Ma. Rutgers University, USA
Tiny Autoencoders are Effective Few-Shot Generative Model Detectors. WIFS, 2024
Luca Bindini, Giulia Bertazzini, Daniele Baracchi, Dasara Shullani, Paolo Frasconi, and Alessandro Piva. University of Florence, Italy
A siamese-based verification system for open-set architecture attribution of synthetic images. PRL, 2024
Lydia Abady, Jun Wang, Benedetta Tondi, Mauro Barni. University of Siena, Italy

Detecting Origin Attribution for Text-to-Image Diffusion Models. WACV, 2025
Katherine Xu, Lingzhi Zhang, Jianbo Shi. University of Pennsylvania, USA
ReTD: Reconstruction-Based Traceability Detection for Generated Images ICASSP, 2025
Weizhuo Chen, Fangfang Yuan, Cong Cao, Kun Peng, Dakui Wang, Yanbing Liu. Chinese Academy of Sciences, China
核心思想：重建残差送入（CNN+Transformer+MLP）训练完成检测
Few-Shot Class-Incremental Model Attribution Using Learnable Representation From CLIP-ViT Features. arXiv, 2025
Hanbyul Lee, Juneho Yi. Sungkyunkwan University, South Korea
LoRAX: LoRA eXpandable Networks for Continual Synthetic Image Attribution arXiv, 2025
Danielle Sullivan-Pao, Nicole Tian, Pooya Khorrami. MIT Lincoln Laboratory & Yale University, USA