Paper 2024/588

Digital Signatures for Authenticating Compressed JPEG Images

Simon Erfurth, University of Southern Denmark

We construct a digital signature scheme for images that allows the image to be compressed without invalidating the signature. More specifically, given a JPEG image signed with our signature scheme, a third party can compress the image using JPEG compression, and, as long as the quantization tables only include powers of two, derive a valid signature for the compressed image, without access to the secret signing key, and without interaction with the signer. Our scheme is constructed using a standard digital signature scheme and a hash function as building blocks. This form of signatures that allow image compression could be useful in mitigating some of the threats posed by generative AI and fake news, without interfering with all uses of generative AI. Taking inspiration from related signature schemes, we define a notion of unforgeability and prove our construction to be secure. Additionally, we show that our signatures have size 32.5 kb under standard parameter choices. Using image quality assessment metrics, we show that JPEG compression with parameters as specified by our scheme, does not result in perceivably reduced visual fidelity, compared to standard JPEG compression.

Available format(s)
Publication info
digital signaturesJPEG Compressionhomomorphic signatures
Contact author(s)
simon @ serfurth dk
2024-04-16: approved
2024-04-16: received
See all versions
Short URL
Creative Commons Attribution


      author = {Simon Erfurth},
      title = {Digital Signatures for Authenticating Compressed JPEG Images},
      howpublished = {Cryptology ePrint Archive, Paper 2024/588},
      year = {2024},
      note = {\url{}},
      url = {}
Note: In order to protect the privacy of readers, does not use cookies or embedded third party content.