![](bg6.png)
Table 1: Image quality using a 7,3 linear error-block code.
Heterogeneity
thresholds
Number of blocks provided
by each bit plane
MSE
PSNR ()
1
1.8 8286 0.1798 55.5816
4
1
4,7
9399,8387)
0.3711 52.4349
43
4,7 9399,8387 0.6487 50.0104
52
4,7
7519,12580)
Not enough 1
s
plane blocks
52
3.5,11.3
8207,8327)
0.2951 53.4308
421
4.7,11,20 8521,8521,8514 0.7438 49.4159
given block can be caused by the alteration of any
number of its bits, which will give the same value
for different number of altered bits within the same
block. Consequently, the reliability of our protocol is
only assessed by the influence of alteration on the
image, presented by the MSE and PSNR values. In
Table 1 we also give the number of blocks provided
by each bit plane to construct carrier vectors. The
classical protocol corresponds to the classical type
1
given in the first row of the table. Note that it
uses only least significant bits. Therefore, to get
enough information carrier bits, we had to set the
heterogeneity threshold to a very low value.
By comparing MSE and PSNR values, it is clear
that partition
4
1
is better than 43. Both of
them use 2 bit planes and require the same number
of bits, but the first partition has a better structure as
was described in Paragraph 5, Section 4. Partition
52 with heterogeneity thresholds 4,7 does not
provide enough bits to embed the message. We
adjusted these to 3.5,11.3 which provided higher
embedding capacity, though at the cost of some
cover quality degradation. Notice that partition
421 as it uses 3 bit planes, provides a large
embedding capacity, which allowed us to set high
heterogeneity thresholds, and obtain close MSE and
PSNR values to the other partitions.
6 CONCLUSIONS
Comparing to classical code based steganography
protocols, our scheme based on linear error-block
codes increases the number of exploitable bits in a
given cover. Specifically, multiple bit planes in an
image are exploited whilst maintaining good MSE
and PSNR values. A major factor to get the
maximum benefit from this scheme is the choice of
the cover, as heterogeneous pixels within an image
promote using multiple bit planes. Heterogeneity
thresholds determine the number of bit planes to use
in each pixel in order to keep good image quality
parameters.
In this paper we compared the results of
embedding using different types of a given code,
including the classical LSB embedding which
corresponds to the classical type. Forthcoming
works involve finding optimal codes to use with this
method. Also, by statistical studies, the list of
heterogeneity thresholds is likely to be optimized.
REFERENCES
Crandall, R., 1998. Some notes on steganography. Posted
on the steganography mailing list.
Westfeld, A., 2001. F5 - a steganographic algorithm. In
Ira S. Moskowitz, editor, Information Hiding, Lecture
Notes in Computer Science, vol. 2137, pp. 289-302.
Springer Berlin Heidelberg.
Zhang W., Zhang, X., Wang, S., 2007. A double layered
"plus-minus one" data embedding scheme. IEEE
Signal Processing Letters, vol. 14, no. 11, pp. 848-
851.
Zhang, X., Zhang, W., Wang, S., 2007. Efficient double-
layered steganographic embedding. Electronics
Letters, vol. 43, no. 8, pp. 482-483.
Liao, X., Wen, Q. Y., 2008. Embedding in two least
significant bits with wet paper coding. In Computer
Science and Software Engineering, vol. 3, pp. 555-
558. IEEE Computer Society.
Dariti, R., Souidi, E. M., 2011. Improving Code-Based
Steganography with Linear Error-Block Codes. In
Digital Information Processing and Communications,
CCIS, vol. 189, pp. 310-321, Springer Berlin
Heidelberg.
Feng, K., Xu, L., Hickernell, F. J., 2006. Linear error-
block codes. Finite Fields and Their Applications, vol.
12, no. 4, pp. 638-652.
Tan, J., Luo, X., Cheng, Q., 2003. A lossless data
embedding scheme for medical in application of e-
diagnosis. In Proceedings of the 25th Annual
International Conference of the IEEE EMBS Cancun.
Dariti, R., Souidi, E. M., 2013. Packing and Covering
Radii of Linear Error-Block Codes. International
Journal of Mathematical, Computational Science and
Engineering, vol. 7, no. 4, pp. 14-18. World Academy
of Science, Engineering and Technology.
ASteganographicProtocolBasedonLinearError-BlockCodes
183