Privacy-preserving inpainting for outsourced image

In this article, a framework of privacy-preserving inpainting for outsourced image and an encrypted-image scheme are proposed. Different with conventional in plaintext domain, there two entities, that is, content owner restorer, our framework. Content first encrypts his or her damaged privacy protection outsources the encrypted, to who may be cloud server powerful computation capability. Image restorer performs encrypted domain sends inpainted back authorized receiver, can acquire final result through decryption. scheme, assist Johnson–Lindenstrauss transform preserve Euclidean distance between vectors before after encryption, best-matching block smallest current found utilized patch filling Paillier-encrypted image. To eliminate mosaic effect decryption, weighted mean filtering is conducted Paillier homomorphic properties. Experimental results show effectively applied secure computing, proposed achieves comparable visual quality some typical schemes domain.