Reversible image hiding scheme using predictive coding and histogram shifting Source: Authors: Reporter: Date: Signal Processing, Vol.89, Issue 6, pp , June 2009 Piyu Tsai, Yu-Chen Hu, Hsiu-Lien Yen Wan-Yu Lu 2013/01/10
2 Outline Introduction Related work - Ni et al.s scheme (The histogram-based reversible data hiding ) The proposed method - Embedding phase - Extraction phase Experimental results Conclusion Comparatives
3 Introduction Embedding algorithm stego-image Secret data Embedded imageOriginal image Extraction algorithm
4 Related work (1/2) peak pixels Peak = 3 Zero = 7 Original image Zero Ni, Zhicheng., Shi, Y.Q., Ansari, N. and Su, Wei. (2006), Reversible Data Hiding, IEEE Transactions on Circuits and Systems for Video Technology, 16 (3), pp
5 Related work (2/2) Secret data = pixels Embedded image Ni, Zhicheng., Shi, Y.Q., Ansari, N. and Su, Wei. (2006), Reversible Data Hiding, IEEE Transactions on Circuits and Systems for Video Technology, 16 (3), pp
6 The proposed method (1/6) error value = basic pixel - residual pixel rrr rbr rrr Embedding phase Original image
7 The proposed method (2/6) Embedding phase peak 1 peak 2 zero 1 zero 2 peak 1 = -1, zero 1 = -4 peak 2 = 0, zero 2 =
8 The proposed method (3/6) Embedding phase Secret data={ }
9 The proposed method (4/6) Embedding phase residual pixel = basic pixel - error value rrr rbr rrr Embedded image
10 The proposed method (5/6) Extraction phase peak 1 = -1, zero 1 = -4 peak 2 = 0, zero 2 = 4 Secret data={ }
11 The proposed method (6/6) Extraction phase
12 Experimental results (1/4)
13 Experimental results (2/4) *4
14 Experimental results (3/4)
15 Experimental results (4/4) Ni et al’s scheme peak=16591 Capacity=16591 bits Proposed scheme peak=23723 peak=13539 Capacity=37262 bits
16 Conclusion To achieve a higher hiding capacity, secret data are embedded in the residual images instead of the image histogram.
17 Comparisons To compare with related work: - Reversible medical image watermarking based on wavelet histogram shifting (2011) Same - Reversible medical image hiding based on histogram shifting. - Two-pairs Difference - spatial domain vs. transformed domain (IDWT) - overflow and underflow
Thanks for your attention !