Cryptography Reference
In-Depth Information
Fig. 13.5.
Binary to L-ary conversion using a variant of arithmetic encoding.
2) Divide R into N equal subintervals, R
0
to R
N−1
.
3) Set R = R
w
, where w = s
w
−Q
L
(s
w
) is the current watermark symbol.
4) If there are remaining symbols, go to Step 1. Find shortest binary string
H∈R.
The classical LSB modification, which embeds a binary symbol (bit) by
overwriting the least significant bit of a signal sample, is a special case. Here
L = 2. G-LSB embedding enables the embedding of a non-integer number of
bits in each signal sample. Thus, it introduces new operating points along the
rate- or capacity-distortion curve.
Lossless Generalized-LSB Data Embedding And Extraction
Fig. 13.6 shows a block diagram of the proposed algorithm. In the embedding
phase, the host signal is quantized and the residual is obtained in Eq. (13.9).
Then they adopt the CALIC lossless image compression algorithm. This has
the quantized values as side information, to e
ciently compress the quanti-
zation residuals in order to create high capacity for the payload data. The
compressed residual and the payload data are concatenated and embedded
into the host signal using the generalized-LSB modification method. The re-
sulting bit stream is converted to L-ary symbols as mentioned above. This
is then added to the quantized host to form the watermarked signal s
w
in
Eq. (13.6). Note that the compression block uses the rest of the host sig-
nal, Q
L
(s), as side-information, to facilitate better compression and higher
capacity.
