Presentation is loading. Please wait.

Presentation is loading. Please wait.

Source: Information Sciences, 2018, accpeted.

Similar presentations


Presentation on theme: "Source: Information Sciences, 2018, accpeted."— Presentation transcript:

1 Reversible data hiding using multi-pass pixel-value-ordering and pairwise prediction-error expansion
Source: Information Sciences, 2018, accpeted. Authors: Wenguang He , Gangqiang Xiong, Shaowei Weng, Zhanchuan Cai, Yaomin Wang Speaker :Xiaozhu Xie Date : /08/09

2 Outline Related work Proposed scheme Experimental results Conclusions
Li et al.’s PVO Peng et al.’s IPVO Qu et al.’s PVO-based pairwise PEE Proposed scheme Experimental results Conclusions I would like to present in these five parts.

3 Related work - Li et al.’s PVO (1/9)
A cover block Li X, Li J, Li B, et al. High-fidelity reversible data hiding scheme based on pixel-value-ordering and prediction-error expansion[J]. Signal processing, 2013, 93(1): 3

4 Related work - Li et al.’s PVO (2/9)
Example of Li et al.’s PVO by modifying the maximum of a block 3

5 Related work-Peng et al.’s IPVO (3/9)
Improved maximum-modification-based data embedding Redefine 𝑑 𝑚𝑎𝑥 𝑑 𝑚𝑎𝑥 = 𝑥 𝜎(𝑛) − 𝑥 𝜎 𝑛−1 >0 if 𝜎 𝑛 <𝜎(𝑛−1) 𝑥 𝜎(𝑛−1) − 𝑥 𝜎 𝑛 ≤0 if 𝜎 𝑛 >𝜎(𝑛−1) 3 [29] F. Peng , X. Li , B. Yang , Improved pvo-based reversible data hiding, Digit. Signal Process. 25 (2014) 255–265 .

6 Related work-Peng et al.’s IPVO (4/9)
Improved maximum-modification-based data embedding 𝑥 𝜎(𝑛) b: the to-be-embedded bit. 3

7 Related work-Peng et al.’s IPVO (5/9)
Improved maximum-modification-based data embedding Example of Peng et al.’s IPVO by modifying the maximum of a block 3

8 Related work-Peng et al.’s IPVO (6/9)
Improved minimum-modification-based data embedding Redefine 𝑑 𝑚𝑖𝑛 Data embedding 𝑑 𝑚𝑖𝑛 = 𝑥 𝜎(1) − 𝑥 𝜎 2 <0 if 𝜎 1 <𝜎(2) 𝑥 𝜎(2) − 𝑥 𝜎 1 ≥0 if 𝜎 1 >𝜎(2) 𝑥 𝜎(1) 3

9 Related work-Ou et al.’s PVO-based pairwise PEE (7/9)
Redefine 𝑒 𝑚𝑎𝑥 , viz., 𝑑 𝑚𝑎𝑥 Modification manner on the 2D-PEH of Peng et al.’s IPVO ( 𝑓 2 ) [22] B. Ou , X. Li , J. Wang , High-fidelity reversible data hiding based on pixel-value-ordering and pairwise prediction-error expansion, J. Vis. Commun. Image R 39 (2016) 12–23 . 3

10 Related work-Ou et al.’s PVO-based pairwise PEE (8/9)
C D E F 2D-PEH generated by Ou et al.’s method on Lena image (2×3 blocksize) 2D mapping for Ou et al.’s PVO-based pairwise PEE( 𝑓 1 ) 3

11 Related work-Ou et al.’s PVO-based pairwise PEE (9/9)
Threshold: 𝑇 1 , 𝑇 2 NL: Noise level 11

12 Proposed scheme (1/6)

13 Proposed scheme (2/6) F B C D A E
High power version of PVO-based pairwise PEE (improve 𝑓 1 ) A B C D E F Redefine 𝑒 𝑚𝑎𝑥 𝑒 𝑚𝑎𝑥 2 = 𝑝 𝑣 − 𝑝 𝜎 𝑛−2 >0 𝑣<𝜎 𝑛−2 𝑝 𝜎 𝑛−2 −𝑝 𝑣 ≤0 𝑣>𝜎 𝑛−2

14 Proposed scheme (3/6) High power version of PVO-based pairwise PEE (improve 𝑓 1 ) Data embedding 2D mapping for the proposed pairwise PEE( 𝑓 𝐻 )

15 Proposed scheme (4/6) Low power version of PVO-based pairwise PEE(improve 𝑓 2 ) ≤1 ≤1 2D mapping for the proposed pairwise PEE( 𝑓 𝐿 )

16 Proposed scheme (5/6) Optimal-threshold-determination
𝑇 1 ∗ , 𝑇 2 ∗ : Two threshold. 𝐷 𝐻 𝑡 , 𝐷 𝐿 𝑡 : Distortion generated by 𝑓 𝐻 , 𝑓 𝐿 . 𝐸 𝐻 𝑡 , 𝐸 𝐿 𝑡 : Capacity generated by 𝑓 𝐻 , 𝑓 𝐿 . EC: Embedding capacity. 60+ 𝐿 𝑐𝑙𝑚 : The length of auxiliary information and compressed location map.

17 Proposed scheme (6/6) Step 1. Location map construction
Blocks with pixel values equal to 0 or 255 are unavailable. Step 2. Payload preparation The LSBs of first 60+ 𝐿 𝑐𝑙𝑚 pixels and secret message make up the real payload. Step 3. Payload embedding For available blocks, calculate 𝑁𝐿 and two error pairs 𝑒 𝑚𝑎𝑥 1 , 𝑒 𝑚𝑎𝑥 2 , 𝑒 𝑚𝑖𝑛 1 , 𝑒 𝑚𝑖𝑛 2 , then embed data guiding by 𝑓. 𝑓= 𝑓 𝐻 𝑁𝐿≤ 𝑇 1 ∗ 𝑓 𝐿 𝑇 1 ∗ <𝑁𝐿≤ 𝑇 2 ∗ Step 4. Auxiliary information embedding

18 Experimental results-(1/8)

19 Experimental results-(2/8)

20 Experimental results-(3/8)
[20] B. Ou , X. Li , Y. Zhao , R. Ni , Reversible data hiding using invariant pixel-value-ordering and prediction-error expansion, Signal Process. Image Commun. 29 (7) (2014) 760–772 . [22] B. Ou , X. Li , J. Wang , High-fidelity reversible data hiding based on pixel-value-ordering and pairwise prediction-error expansion, J. Vis. Commun. Image R 39 (2016) 12–23 . [29] F. Peng , X. Li , B. Yang , Improved pvo-based reversible data hiding, Digit. Signal Process. 25 (2014) 255–265 . [30] X. Qu , H.J. Kim , Pixel-based pixel value ordering predictor for high-fidelity reversible data hiding, Signal Process. 111 (2015) 249–260 . [32] V. Sachnev , H.J. Kim , J. Nam , S. Suresh , Y.Q. Shi , Reversible watermarking algorithm using sorting and prediction, IEEE Trans. Circuits Syst. Video Technol. 19 (7) (2009) 989–999 . Performance comparison between the proposed scheme and other related schemes.

21 Experimental results-(4/8)
Performance comparison between the proposed scheme and other related schemes.

22 Experimental results-(5/8)
Performance comparison between the proposed scheme and other related schemes.

23 Experimental results-(6/8)

24 Experimental results-(7/8)

25 Experimental results-(8/8)

26 CONCLUSIONS Multi-pass PVO prediction.
The high and low power version of PVO-based pairwise PEE. Efficient capacity-distortion trade-off.

27 Thank you!


Download ppt "Source: Information Sciences, 2018, accpeted."

Similar presentations


Ads by Google