Cryptography Reference
In-Depth Information
decoding pattern. For example ensure that the top-left corner is within
the top-left block NN of the locating decoding pattern.
At each translation, evaluate S =
n
−b
i
) for the given image us-
ing the patches coordinates defined by the locating decoding pattern. In
this expression, n represents the number of the pixel values selected for
patch A or B during the embedding process. The values a
i
and b
i
respec-
tively indicate the pixel values of a point pair in the given image. The
same point pair in the locating decoding pattern is a patch pair whose
values are respectively increased and decreased by δ during embedding.
Determine the translation that produces the maximal S, and denote by
(x
(a
i
i=1
′
′
′
′
1
) indicate the
deviation from the top-left pixel of the locating decoding pattern pixel.
Ensure that the retrieved coordinates are in the correct cropped position.
By (x
1
,y
1
) the translation parameters. The coordinates (x
1
,y
1
), we obtain the remnant integral blocks. That is, every block
of NN size.
′
1
,y
′
Fig. 13.14.
The classification of pixel pairs during extraction.
Stage 2: Decoding and Authentication
For each integral block obtained, the remnant pixels are grouped into pairs
of pixel values in each block. We then apply the integer transform given
in Eq. (13.12) to each pair to calculate their average value l and difference
value h. We use the same scanning order as that in embedding to place
all difference values in a one dimensional list
h
1
, h
2
,, h
n
.
