Cryptography Reference
In-Depth Information
file size is because the embedding operation increases the randomness of the
image data to some degree, and the lossless compression by LHA algorithm
is not as effective as it is for non-embedded images. The embedding ratio is
very high in all embeddings. Especially, in PNG, we knew that the embedding
ratio were almost always 0.65 or around. This is a very high score. Before and
after embedding differences are shown in Fig. 8.21 (A) and (B) in black and
white image patterns. This is about the Red color component. Black area
in this pattern indicates the unaltered area in embedding, while gray area
shows the altered area. The whiter the area, the larger the alteration was.
The maximum difference between before and after was about 30 in decimal
number (out of 256 color values each). Green and blue colors were almost the
same as Red color. Embedding speed was quite high. In case of a notebook
PC having a 1.4GH AMD Athlon CPU with 384MB memory, each embedding
time in BMP image is shown in Table 8.5. PNG output took a little more time
(0.2 second or less) in both cases.
Table 8.5.
Properties of the stego images (RGB color).
Image name Image size BMP data PNG data Embedding ratio
Embedding
size (KB)
size (KB) BMP
PNG
time (BMP)
Image-1
950713
1,986
1,584
0.52
0.65
2.0 sec (approx)
Image-2
600465
818
634
0.43
0.66
1.0 sec (approx)
(A) Image difference in Red color
(B) Image difference in Red color
(Image-1)
(Image-2)
Fig. 8.21.
Image difference patterns between before and after BPCS embedding in
Red color component.
Extracting
The operation in the extracting program starts by checking the access key
