PRISM: Phase-enhanced Radial-based Image Signature Mapping framework for fingerprinting AI-generated images

2025-09-18

AI Summary

To address the untraceable provenance of generative AI content—particularly critical for accountability in commercial, paid applications—this paper proposes a lightweight, model-agnostic image fingerprinting framework. Our method innovatively fuses Fourier amplitude and phase information, augmented by radial frequency compression, phase enhancement, and Linear Discriminant Analysis (LDA), enabling robust cross-architecture and cross-dataset identification of generative models. Evaluated on our large-scale, self-constructed dataset PRISM-36K, the framework achieves 92.04% attribution accuracy; across four standard benchmarks, it attains a mean accuracy of 81.60%, and improves binary real/fake image classification to 95.06%. Unlike prior approaches requiring model access or retraining, our method operates solely on generated images, ensuring broad deployability. It significantly strengthens content traceability and system trustworthiness, providing a scalable, practical foundation for AI content governance and provenance verification.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

A critical need has emerged for generative AI: attribution methods. That is, solutions that can identify the model originating AI-generated content. This feature, generally relevant in multimodal applications, is especially sensitive in commercial settings where users subscribe to paid proprietary services and expect guarantees about the source of the content they receive. To address these issues, we introduce PRISM, a scalable Phase-enhanced Radial-based Image Signature Mapping framework for fingerprinting AI-generated images. PRISM is based on a radial reduction of the discrete Fourier transform that leverages amplitude and phase information to capture model-specific signatures. The output of the above process is subsequently clustered via linear discriminant analysis to achieve reliable model attribution in diverse settings, even if the model's internal details are inaccessible. To support our work, we construct PRISM-36K, a novel dataset of 36,000 images generated by six text-to-image GAN- and diffusion-based models. On this dataset, PRISM achieves an attribution accuracy of 92.04%. We additionally evaluate our method on four benchmarks from the literature, reaching an average accuracy of 81.60%. Finally, we evaluate our methodology also in the binary task of detecting real vs fake images, achieving an average accuracy of 88.41%. We obtain our best result on GenImage with an accuracy of 95.06%, whereas the original benchmark achieved 82.20%. Our results demonstrate the effectiveness of frequency-domain fingerprinting for cross-architecture and cross-dataset model attribution, offering a viable solution for enforcing accountability and trust in generative AI systems.

Problem

• Identifying model sources for AI-generated images
• Fingerprinting images using frequency-domain signatures
• Ensuring accountability in commercial generative AI systems

Innovation

• Phase-enhanced radial Fourier transform fingerprinting
• Linear discriminant analysis for signature clustering
• Frequency-domain signatures for cross-architecture attribution