Fast Deterministic Self-Blinding for Metadata-Free Integrity Verification of Paillier Encrypted Data: Application to 3D Models

Ensuring the integrity and authenticity of homomorphically encrypted data without relying on auxiliary metadata remains a fundamental challenge in secure outsourced computation. Existing methods based on the probabilistic self-blinding (PSB) property of the Paillier cryptosystem enable only fragile watermarking, require costly random trials, and cannot embed a cryptographic signature directly into a ciphertext. Consequently, they still depend on external metadata to guarantee full integrity. This work introduces a deterministic self-blinding (DSB) mechanism that, for the first time, enables efficient and constant time modulation of Paillier ciphertexts while preserving both the underlying plaintext and the IND-CPA security of the scheme. The determinism of DSB, combined with Paillier’s homomorphic structure, enables embedding a complete cryptographic signature of the ciphertext inside itself, thereby providing self-contained, metadata-free integrity and authenticity verification entirely in the encrypted domain. We further incorporate quantization index modulation (QIM) to support robust traceability after decryption. Experiments on 3D models, chosen as a representative use case for high-volume data, demonstrate that the proposed scheme maintains high visual quality, achieves an embedding speedup of approximately 10 5 over PSB approaches, and provides complete integrity/authenticity control applicable to any Paillier-encrypted data. These results position DSB as an efficient and secure foundation for encrypted-domain processing in untrusted cloud environments.