Image Secret Sharing - Visual Cryptography and Secret Image Sharing - page 391

FIGURE 14.11
(See color insert.) Breaking the correlation of neighboring blocks in an image.

nique. What is different here from the earlier examples (where we created a
whole new image as a share) is that we embed the information of each rectan-
gular block into one of the image shares. The embedded information of share
1 is in the lower-left corner, that of share 2 is in the upper-left position while
that of share 3 is embedded is in the upper-right region, and that of share 4
is in the lower-right corner. Thus, the information of the original secret image
is shared among all the image shares. While this example is a simple illustra-
tion of how the correlation can be broken, it is clear that more sophisticated
secret image subdivision and sharing schemes can be devised using the same
principle.

14.5 Conclusion
In this chapter, we introduce the novel concept of color image sharing based on
Lagrange interpolation and moving lines. The security of secret sharing is built
upon the computation of polynomials. Given enough shares, the coecients
of the interpolating polynomial can be exactly determined. However, if an
attacker does not possess the threshold number of shares and the image shares,
he will not be able to reconstruct the polynomial. Thus, he will not able to
restore the secret image properly.
Secret sharing based on Lagrange interpolation is often utilized to share
binary strings, but it is dicult to use for color image sharing since it yields
only a limited number of shares. We therefore have developed a new color im-
age sharing scheme based on moving lines that does not have that limitation.
An implicit curve generated by moving lines is a rational curve; we believe it
can therefore be directly applied for sharing compressed-domain images.
Our future work will focus on further investigation of color image sharing
in the compressed domain. The possible research directions in the future are:

1. Color image sharing based on compressed domain processing. Ideas

related to DCT, DFT, and DWT can be employed in color image
sharing.