Cryptography Reference
In-Depth Information
Step 7. Perform Du et al.s method to extract the serial number Y using
all image data X(i, j) satisfying thatf (c X(i,j) )> 1.
Step 8. Transform the serial number Y into a bit stream. Then, combine
these two bit streams as the secret data A to be the output.
5.4 Experimental Results
In this chapter, we present some experiments and their results to help evalu-
ate the performance of the proposed method. In our experiments, six palette
images were used as test images. These are shown in Fig. 5.10. These images
are at most 256 different colors. Fig. 5.10(a) to (d) show the downloaded im-
ages from several homepages on the Internet. Fig. 5.10(e) and Fig. 5.10(f) are
of comic pictures.
To evaluate the image quality of a processed image of hw, the peak-
signal-noise ratio (PSNR) was used. The formula of PSNR is defined as follows:
255 2
MSE
PSNR = 10 log 10
(dB),
(5.11)
h−1
w−1
3
1
hw
−β i,j,k ) 2 ,
MSE =
i,j,k
(5.12)
i=0
j=0
k=1
where α i,j and β i,j denote the original and compressed color pixels, respec-
tively. Typically, a larger PSNR value stands for a better image quality.
Table 5.1 shows how Tzeng et al.s method and our proposed method
compare when the same thresholds T α and T β (T α =1andT β = 30) are used
to embed the secret data. As shown in the results, our proposed method can
offer a bigger embedding capacity than Tzeng et al.s, and the image quality
our proposed method provides is either better or about the same as Tzeng et
al.s.
The Stego-Images are shown in Fig. 5.11. According to our experimental
results, using our proposed method, the secret data can be embedded into the
cover image without introducing visual artifacts. The secret data embedded
in the stego-image can be extracted correctly.
Furthermore, Table 5.2 confirms the high embedding capacity and good
image quality that our proposed method is able to provide. Those experimen-
tal results were based on the fixed threshold values T α and T β (T α =1and
T β = 30) and a variable threshold value of T γ . According to the threshold
value T γ , the maximum merged clusters could be obtained by our clustering
scheme, and those results could be used to embed secret data. Therefore, the
threshold value T γ was set as 10, 15,, 50, and the resultant cluster sizes,
embedding capacities, and PSNR values obtained are shown in Table 5.2.
Fig. 5.12(a) shows the relationship between the number of clusters and
the corresponding embedding capacity. Fig. 5.12(b) shows the relationship
Search WWH ::




Custom Search