Cryptography Reference
In-Depth Information
Fig. 5.1.
Flowchart of data-hiding scheme.
Association, for example [5]. Several image hiding techniques have been de-
veloped for gray-scale as well as color images [6, 7, 8, 9], but less attention
was paid to applications for Palette Images.
Palette-based images, or simply called Palette Images, are ubiquitous in
modern computer systems. For example, GIF and 8-bit BMP files are some
palette images widely used in multimedia applications. Each palette image
consists of two components: a fixed lookup table called the Color Palette and
a series of color indices called Image Data. Since image data in a palette image
are not color values but rather indices of the color palette, it means that hiding
data in a palette image is more challenging than in the pixel values of an image.
We have two possibilities: the secret information can be embedded either
in the color palette itself or in the image data. In the literature concerned,
not much information can be found on palette image data hiding. In 1998,
Kwan [10] proposed a program called Gifshu
e to embed data in GIF
images. The idea is that the secret data used to permute the colors on the
palette of an image in some specific order. For the 256 entries of a color
palette, there are 256! possible permutations. So, at most log
2
(256!) bits can
be embedded into a GIF image. The advantage of this method is that the
visual content of the cover image is not affected. However, security is weak
because when the image is reloaded or saved, the image-processing software,
whatever it may be, will usually rearrange the order of the colors on the
palette according to some factors such as luminance, frequency of occurrence,
for example. In 1999, Fridrich and Du proposed a method [11] to embed data
into the parity bits of closest colors. The method first assigns a specific parity
bit (0 or 1) to each color on the palette, and then modulates the index values
to make the parities of the new index values replaced by the closest color
equal to the bits of the embedded data. This adaptive method can embed
an appropriate amount of data by modification of the pixel values. In 2004,
Chang, Tsai and Lin proposed a method [12] to hide secret data based on the
Color Clustering Technique. The colors in the palette are grouped into sub-
clusters according to the relationship between the colors. For each image data,
the hiding capacity is determined by the size of the sub-cluster of which it is a
member. This method embeds a considerable quantity of data and has a good
