Decentralizing Photo Forensics for Public and Verifiable Trust

Abstract

The integrity and traceability of digital photographic evidence represent a critical factor during forensic investigations, especially when this evidence undergoes technical transformations, such as cropping or resolution enhancement. Ensuring that these modifications remain transparent, verifiable, and attributable is essential to maintaining the value of the evidence during an investigation. To meet these requirements, existing systems typically rely on blockchain-based implementations within permissioned networks or provide only limited support for image transformations. As a result, they often lack the flexibility and transparency required for open or decentralized forensic scenarios. In this paper, we propose an endorsement-based image forensics system that leverages public blockchain to record the lifecycle and verify the authenticity of images. Our system employs hybrid encryption to provide confidentiality of uploaded images while simultaneously ensuring that they remain auditable and non-repudiable. The system supports different trust models and enables users to assess the trustworthiness of an image’s provenance data directly and indirectly. Direct trust is achieved by validating an image transformation through reproducible functions or zero-knowledge proofs; indirect trust is enabled through publicly recorded endorsements. Our design achieves low gas costs and provides confidentiality, verifiability, and traceability guarantees, improving upon previous approaches without relying on permissioned infrastructures.