Cryptography Reference
In-Depth Information
map are highly correlated, enough space could be provided by compressing
the LSB of the segment containing 128 vertices. However, as the scheme can
not be practically applied as the Huffman code book was not considered.
6.5 Summary and Future Works
The Digital Vector Map is an important part of geographical information sys-
tem(GIS) and has been widely used in many regions. In this chapter the topic
of GIS watermarking has been considered. That is, the techniques of hiding
data in 2D vector maps. This has many potential usages in the applications of
vector map data. According to the properties and the purpose of the schemes,
related works in this field were categorized into three classes. That is, robust
watermarking, reversible data hiding and fragile watermarking. The bound-
aries among these classes are not absolutely distinct. For example, the fragile
scheme presented in Section 6.4 is also an invertible scheme. The state of the
art of the 2D vector map data hiding was given in this chapter by introducing
the main proposed schemes in each class. Also some open issues in current
study were presented several schemes. Some drawbacks of former works were
improved in the schemes presented. By investigating the performance of the
current schemes, it is shown that there are still many unsolved problems.
The robustness to map simplification and additive noise attacks needs to be
further strengthened. Secondly, the distortions introduced by watermarking
should be controlled in multi-aspects. Thirdly, the evaluation of map quality
needs an appropriate measure of performance. Furthermore, reversible and
fragile schemes need further consideration because of the high security level
necessary in some special applications.
6.5.1 Robustness to Certain Attacks
Most proposed algorithms have good performance to resist geometrical at-
tacks, but are less robust against additive noise attack. Most schemes are
fragile to map simplification or cropping. This is especially for attacks im-
plemented in the transform domain. It is understandable that the more in-
formation lost during the attack, the more dangerous it is. Maliciously im-
posing noise on map data is not a good choice for an attacker, small noise
distortions do occur in some daily works such as map format transforma-
tion. Consequently, a robust scheme should at least be robust to noises with
low amplitudes. Map simplification is a common operation in GIS and can
cause information lost without degrading the validity of the map data. It is
very important for a watermarking scheme to resist this attack. This can be
considered an open issue and needs to be enhanced.
