ZK-WAGON: Imperceptible Watermark for Image Generation Models using ZK-SNARKs

View on arXiv ← Back to list

Authors: Aadarsh Anantha Ramakrishnan, Shubham Agarwal, Selvanayagam S, Kunwar Singh

Published: 2025-10-02 12:39:57+00:00

AI Summary

ZK-WAGON is a novel system that uses Zero-Knowledge Succinct Non-Interactive Arguments of Knowledge (ZK-SNARKs) to create imperceptible, verifiable watermarks for images generated by both GAN and Diffusion models. The system generates a cryptographic proof of origin without exposing proprietary model weights, generation prompts, or sensitive internal information. This proof is then imperceptibly embedded into the generated image using Least Significant Bit (LSB) steganography, providing a secure and model-agnostic verification pipeline.

Abstract

As image generation models grow increasingly powerful and accessible, concerns around authenticity, ownership, and misuse of synthetic media have become critical. The ability to generate lifelike images indistinguishable from real ones introduces risks such as misinformation, deepfakes, and intellectual property violations. Traditional watermarking methods either degrade image quality, are easily removed, or require access to confidential model internals - making them unsuitable for secure and scalable deployment. We are the first to introduce ZK-WAGON, a novel system for watermarking image generation models using the Zero-Knowledge Succinct Non Interactive Argument of Knowledge (ZK-SNARKs). Our approach enables verifiable proof of origin without exposing model weights, generation prompts, or any sensitive internal information. We propose Selective Layer ZK-Circuit Creation (SL-ZKCC), a method to selectively convert key layers of an image generation model into a circuit, reducing proof generation time significantly. Generated ZK-SNARK proofs are imperceptibly embedded into a generated image via Least Significant Bit (LSB) steganography. We demonstrate this system on both GAN and Diffusion models, providing a secure, model-agnostic pipeline for trustworthy AI image generation.

Key findings

The ZK-WAGON framework provides a secure, model-agnostic, and end-to-end verifiable pipeline for trustworthy AI image generation, demonstrated successfully on both GAN and Diffusion architectures. The use of SL-ZKCC drastically reduces the computational complexity and time required for cryptographic proof generation. The resulting watermark is visually imperceptible, ensuring post-generation authenticity without image quality degradation, while allowing users to verify the image's origin without access to proprietary model data.

Approach

The core method is Selective Layer ZK-Circuit Creation (SL-ZKCC), which converts only key layers of the image generation model (like TinyGAN or Stable Diffusion) into a ZK-SNARK circuit to minimize proof generation time. The resulting ZK-SNARK proof, guaranteeing cryptographic origin, is compressed and then embedded into the generated image's pixel data using LSB steganography. Verification involves extracting and decompressing the proof, recomputing a perceptual hash for tamper verification, and checking the ZK-SNARK against the public verification key.

Datasets

UNKNOWN

Model(s)

TinyGAN, Stable Diffusion 2.1 Base (modified with TAESD), Halo2 (proving system)

Author countries

India

← Previous