Fractal Signatures - Securing AI-Generated Pollock-Style Art via Intrinsic Watermarking and Blockchain

Oct 28, 2026

Upload by James

The digital art market faces authenticity verification and copyright protection challenges as AI-generated artworks scale. This study proposes an integrated framework that combines neural style transfer, fractal analysis, and blockchain technology to generate Pollock-inspired abstract art and embed robust, imperceptible intrinsic watermarks derived from fractal and turbulence features. The watermark is secured by binding it to NFT metadata to enable immutable ownership proof. Experiments report a 76.2% average detection rate under common attacks, outperforming traditional 27.8–44.0% methods, improving trust in the digital art ecosystem.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages19

Reading Time48 min

AccessFree to download

Abstract

1 Introduction

2 Related Works

2.1 AI in Art Creation and the Authenticity Gap

3 Methodology

4 Experimental Results

5 Discussion

6 Conclusion

arXiv :2410 .20519v4 [ cs .CV] 23 Jul 2025 Fractal Signatures: Securing AI-Generated Pollock-Style Art via Intrinsic Watermarking and Blockchain Yiquan Wang College of Mathematics and System Science, Xinjiang University, Urumqi, Xinjiang, 830046, China [ethan@stu.xju.edu.cn](ethan@stu.xju.edu.cn) Abstract The digital art market faces unprecedented challenges in authenticity verification and copyright protection. This study introduces an integrated framework to address these issues by combining neural style transfer, fractal analysis, and blockchain technology. We generate abstract artworks inspired by Jackson Pollock, using their inherent mathematical complexity to create robust, imperceptible watermarks. Our method embeds these watermarks, derived from fractal and turbulence features, directly into the artwork’s structure. This approach is then secured by linking the watermark to NFT metadata, ensuring immutable proof of ownership. Rigorous testing shows our feature-based watermarking achieves a 76.2% average detection rate against common attacks, significantly outperforming traditional methods (27.8-44.0%) . This work offers a practical solution for digital artists and collectors, enhancing security and trust in the digital art ecosystem. 1 Introduction Artificial intelligence has fundamentally transformed digital art, changing how artworks are conceived, generated, and distributed in the field of computational creativity. Recent advances in deep learning, particularly Generative Adversarial Networks (GANs) and neural style transfer techniques, have enabled unprecedented capabilities in automated art generation, allowing computers to emulate complex artistic styles and create visually compelling digital artworks [1–3] . Simultaneously, the rise of blockchain technology and Non-Fungible Tokens (NFTs) has revolutionized digital art markets, providing new mechanisms for establishing provenance, ownership, and authenticity of digital creative works [4–6] . However, the proliferation of AI-generated digital art presents major challenges in copyright protection and authenticity verification. Traditional digital watermarking techniques, while effective for conventional digital media, often fail because they rely on perceptual masking in predictable regions of an image, a concept that is ill-defined in the chaotic, multi-layered textures of abstract art like Pollock’s [7, 8] . Furthermore, existing approaches typically embed generic watermarks that are not intrinsically linked to the artistic characteristics of the generated content, making them vulnerable to removal or falsification. The integration of AI art generation with robust copyright protection mechanisms remains an underexplored area, particularly for abstract art styles that exhibit complex mathematical properties. This study addresses two key challenges: generating mathematically authentic abstract artworks and embedding robust, imperceptible watermarks for copyright protection by leveraging the artwork’s inherent characteristics. Specifically, we focus on Jackson Pollock’s drip paintings, which have been scientifically proven to exhibit fractal properties [9, 10], providing a unique opportunity to integrate mathematical complexity measures with watermarking techniques. We propose an integrated framework that combines neural style transfer, multifractal analysis, and blockchainbased authentication to address these challenges. Our approach generates Pollock-style artworks using a neural style transfer model while simultaneously extracting fractal and turbulence characteristics that serve as intrinsic watermark features. Unlike conventional watermarking methods that impose external modifications, our technique leverages the mathematical properties inherent in the generated artworks, creating watermarks that are naturally resilient to common digital attacks. The main contributions of this work are threefold: (1) We develop an integrated art generat

FAQs

How does blockchain and NFT metadata contribute to ownership proof?

The method links the watermark to NFT metadata so the ownership/provenance proof remains immutable for verification.

Oct 28, 2026

Upload by James