JPEGs Just Got Snipped: Croppable Signatures Against Deepfake Images

2025-12-01

AI Summary

To address information distortion caused by the proliferation of deepfake images, this paper proposes a cropping-robust image signature scheme that remains verifiable under arbitrary cropping while immediately invalidating semantic manipulations such as deepfakes. Methodologically, we innovatively integrate BLS short signatures with the JPEG compression standard, embedding cryptographic hashes and signatures directly into the DCT domain. This design yields signatures of constant size (O(1)), eliminates the need for private-key re-signing upon cropping, and does not rely on trusted knowledge of cropping locations. Experiments demonstrate that the signature incurs only moderate increases in image file size and maintains high verification rates across diverse cropping scenarios. Consequently, the proposed scheme significantly enhances image source traceability and authenticity assurance.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Search and Optimization: Distributed Search

Application Category

• Security and Privacy: Applications of cryptography
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Deepfakes are a type of synthetic media created using artificial intelligence, specifically deep learning algorithms. This technology can for example superimpose faces and voices onto videos, creating hyper-realistic but artificial representations. Deepfakes pose significant risks regarding misinformation and fake news, because they can spread false information by depicting public figures saying or doing things they never did, undermining public trust. In this paper, we propose a method that leverages BLS signatures (Boneh, Lynn, and Shacham 2004) to implement signatures that remain valid after image cropping, but are invalidated in all the other types of manipulation, including deepfake creation. Our approach does not require who crops the image to know the signature private key or to be trusted in general, and it is O(1) in terms of signature size, making it a practical solution for scenarios where images are disseminated through web servers and cropping is the primary transformation. Finally, we adapted the signature scheme for the JPEG standard, and we experimentally tested the size of a signed image.

Problem

• Develops croppable signatures to authenticate images against deepfakes.
• Ensures signatures remain valid after cropping but not other manipulations.
• Adapts the method for JPEG and tests practical signature size.

Innovation

• BLS signatures remain valid after cropping
• Signature size is constant O(1) for efficiency
• Adapted scheme for JPEG standard compatibility