To address the tripartite threats of confidentiality compromise, unauthorized redistribution, and malicious tampering in cloud-based media sharing, this paper proposes a Verifiable Client-Side Watermarking model for Thumbnail-Preserving Encryption. This solution integrates secure preview, infringement tracing, and tamper recovery into a unified framework. The method employs a pairwise pixel difference perturbation algorithm to achieve high-quality thumbnail preservation. Crucially, a novel global holographic self-embedding mechanism generates an image micro-summary as a recovery watermark prior to encryption, establishing a blind authentication system. During decryption, specific user identities are embedded as invisible watermarks. In cases of tampering, the system utilizes residual redundant information to precisely locate and reconstruct altered regions via a majority voting mechanism without requiring the original image. Experimental results demonstrate the model’s efficacy in providing robust copyright tracking and high-quality tamper recovery. This approach significantly expands client-side watermarking capabilities, shifting from singular infringement tracing to comprehensive content authenticity and integrity protection for trustworthy digital media distribution.
Li et al. (Wed,) studied this question.