Cryptography Reference
In-Depth Information
Experiment Results
Some experiments are done to test the performance of the proposed algorithm.
The adopted hash function is MD5. It is a one-way function which can take
an arbitrary length input and return a signature with the fixed length of 128
bits. Fig. 6.16 shows the results of the experiment. The differences between
the original map [Fig. 6.16(a)] and the watermarked map [Fig. 6.16(b)] are
well controlled. If there has been no tampering, the original vector map can
be exactly restored during the extraction procedure. To check the schemes
localization ability, we tampered several coordinates of an arbitrary segment
in the watermarked map. Fig. 6.16(c) is the restored map of the tampered
watermarked map. The tampered segment can be reliably detected and has
been marked in the figure. The other segments are restored without loss.
(a) Original map
(b) Watermarked map
(c) Authenticated map
Fig. 6.16.
Experiment results.
The scheme is fully fragile, where authentication will definitely fail even if
there is only a single bit of the protected data has been altered. Note that the
tamper locating precision of the scheme is limited to the length of the map
segments. The selection of the segment length should ensure that every seg-
ment can provide enough space for inserting the authenticating information.
This is highly related to the type of the cover map and the method of loss-
less compression. For maps with highly correlated coordinates, the segment
length could be selected shorter than that in the maps with lower correlation.
The type of the cover map could have an important influence on the precision
of the schemes tamper locating ability. The lossless compression algorithm
adopted in the experiment is the Huffman coding algorithm which is popu-
larly used for data compression. Since the coordinates of the original contour
