An efficient highcapacity reversible data hiding scheme for

An efficient high-capacity reversible data hiding scheme for encrypted images Source: Journal of Visual Communication and Image Representation Volume 66, January 2020, pp. 1 -12 Authors: Bo Guan, Dawen Xu Speaker: Guanlong Li Date: 2020/3/19 1

Outline • Introduction • Proposed Scheme • Experimental Results • Conclusions 2

Introduction(1/1) Cloud (Add Info) Encrypt key Data hiding key Content owner Receiver 3

Proposed Scheme(1/5) 4

Proposed Scheme(2/5) 1 2 3 … Location Map Generation W 1 2 3 126 127 128 … H 127 128 129 Cover Image … … 126 126 127 128 127 129 … … … Prediction Image 5

Proposed Scheme(3/5) Two-layer MSB 126 127 128 129 128 129 … … … Cover Image 0 0 0 0 0 … … … Location Map 126 126 127 128 127 129 … … … Prediction Image 6

Proposed Scheme(4/5) Image Encryption 126 127 128 01001011010…… 127 128 129 … 53 … 13 254 62 105 185 85 2 48 52 … 63 … … Cover Image Encrypted Image … 7

Proposed Scheme(5/5) Data Embedding Secret 53 13 254 62 105 185 85 2 48 52 … 63 11010010…… … … Encrypted Image 0 0 0 0 0 … 245 77 62 190 94 62 255 73 197 63 143 … … Encrypted Image containing hidden data … Location Map(reference) 8

Experimental Results(1/3) 9

Experimental Results(2/3) 10

Experimental Results(3/3) [19] D. W. Xu, R. D. Wang, Separable and error-free reversible data hiding in encrypted images, Signal Process. 123 (2016) 9– 21. [12] Z. Qian, X. Zhang, Reversible data hiding in encrypted images with distributed source encoding, IEEE Trans. Circuits Syst. Video Technol. 26 (4) (2016) 636– 646. [13] P. Puteaux, W. Puech, An efficient MSB prediction-based method for high-capacity reversible data hiding in encrypted images, IEEE Trans. Inf. Forensics Secur. 13 (7) (2018) 1670– 1681. 11

Conclusions • It proposed an efficient high-capacity reversible data hiding method in encrypted images based on MSB prediction, which fully exploits the correlation between neighboring pixels in plaintext image. • Experimental results show that the proposed method can achieve higher embedding capacity compared to most related methods. 12