1 An Efficient VQ-based Data Hiding Scheme Using Voronoi Clustering Authors:Ming-Ni Wu, Puu-An Juang, and Yu-Chiang Li
2 Outline Introduction Related Works Vector Quantization (VQ) Principle Component Analysis (PCA) Proposed method Experimental results Conclusions
3 Introduction The concept of hiding data in a compressed image compress secret information cover image (original image) compressed image compressed stego-image embed produce
4 Vector Quantization (VQ) An image is separated into a set of input vectors Each input vector is matched with a codeword of the codebook
5 Vector Quantization (VQ) Image Index table codebook 0 1 2(132, 21, …,76)
6 Principle Component Analysis (PCA) PCA is a data engineering method to reduce to the complexity of data analysis by projecting values from high dimensional vectors to low dimensional vectors. The projection result on the new subspace will keep the characteristics of the original information.
7 Proposed method Block diagram of codeword clustering phase
8 Proposed method VD construction VD clustering PCA codebook c0c1c2c3c4c5c6c7c0c1c2c3c4c5c6c7 l dimensions p0p1p2p3p4p5p6p7p0p1p2p3p4p5p6p7 2 dimensions
9 Proposed method VD clustering
10 Proposed method Data Embedding Procedure search PCA a secret bit pkpk k cover image index table I x i y i zizi pjpj
11 Proposed method Example 1 for data embedding procedure
12 Proposed method Example 2 for data embedding procedure
13 Experimental results The six cover images
14 Experimental results Table 1. IMAGE QUALITY (PSNR) COMPARISONS OF FOUR SCHEMES WITH SECRET BITS ImagesVQMGLEDHPCAACEDHVC Lena Airplane Boat Barbara Peppers Toys
15 Experimental results Table 2. IMAGE QUALITY (PSNR) COMPARISONS UNDER VARIOUS CAPACITIES WITH IMAGE “ LENA ” CapacityMGLEDHPCAACEDHVC 16K K K K
16 Experimental results (a) VQ compressed(b) DHPCA (c)ACE(d) DHVC
17 Conclusions A novel data hiding scheme which combines PCA and VD techniques The image quality of stego-images is better than previous works