RevHashNet: Perceptually de-hashing real-valued image hashes for similarity retrieval

作者：

Highlights：

•

• We propose the perceptual image de-hashing concept.

• Image de-hashing is shown feasible on several real-valued image hashes.

• The proposed RevHashNet is a deep learning based approach.

• RevHashNet can reconstruct perceptually similar images with high visual quality.

摘要

•We propose the perceptual image de-hashing concept.•Image de-hashing is shown feasible on several real-valued image hashes.•The proposed RevHashNet is a deep learning based approach.•RevHashNet can reconstruct perceptually similar images with high visual quality.

论文关键词：RevHashNet,Image hashing reversion,Image de-hashing,Security,Deep learning

论文评审过程：Received 2 November 2017, Revised 27 June 2018, Accepted 28 June 2018, Available online 30 June 2018, Version of Record 12 July 2018.

论文官网地址：https://doi.org/10.1016/j.image.2018.06.018