ZK-IMG: Attested Images via Zero-Knowledge Proofs to Fight Disinformation

View on arXiv ← Back to list

Authors: Daniel Kang, Tatsunori Hashimoto, Ion Stoica, Yi Sun

Published: 2022-11-09 10:02:20+00:00

AI Summary

ZK-IMG is a novel library that leverages Zero-Knowledge SNARKs (ZK-SNARKs) to attest to image transformations while preserving the privacy of pre-transformed images. It enables the secure and private chaining of arbitrary image transformations on HD images using commodity hardware, overcoming limitations of prior work in terms of privacy, speed, and image resolution. The system compiles high-level image transformation specifications into ZK-SNARKs and uses hashes to hide input and output images during chaining.

Abstract

Over the past few years, AI methods of generating images have been increasing in capabilities, with recent breakthroughs enabling high-resolution, photorealistic deepfakes (artificially generated images with the purpose of misinformation or harm). The rise of deepfakes has potential for social disruption. Recent work has proposed using ZK-SNARKs (zero-knowledge succinct non-interactive argument of knowledge) and attested cameras to verify that images were taken by a camera. ZK-SNARKs allow verification of image transformations non-interactively (i.e., post-hoc) with only standard cryptographic hardness assumptions. Unfortunately, this work does not preserve input privacy, is impractically slow (working only on 128$\\times$128 images), and/or requires custom cryptographic arguments. To address these issues, we present zk-img, a library for attesting to image transformations while hiding the pre-transformed image. zk-img allows application developers to specify high level image transformations. Then, zk-img will transparently compile these specifications to ZK-SNARKs. To hide the input or output images, zk-img will compute the hash of the images inside the ZK-SNARK. We further propose methods of chaining image transformations securely and privately, which allows for arbitrarily many transformations. By combining these optimizations, zk-img is the first system to be able to transform HD images on commodity hardware, securely and privately.

Key findings
ZK-IMG successfully attests to HD image transformations with verification times as low as 5.6 milliseconds. While proving times can be substantial (up to 2236 seconds) and require high peak memory (up to 308 GB) when hashing inputs and outputs for privacy, these operations are feasible on powerful commodity hardware. The system significantly outperforms prior work like PhotoProof, achieving 112x faster proving and 94x faster verification for 128x128 images.
Approach
The authors developed ZK-IMG, a library that transparently compiles high-level image transformation specifications into ZK-SNARKs using the Halo2 proving system and Plonkish arithmetization. To preserve privacy and enable chaining, ZK-IMG computes and reveals only the hashes of input and intermediate images within the ZK-SNARK proofs, revealing only the final transformed image. This allows for arbitrary transformations to be securely and privately applied to HD images.
Datasets
UNKNOWN
Model(s)
ZK-SNARKs, Halo2 proving system, Plonkish arithmetization, Poseidon hash function
Author countries
USA
← Previous
Next →