Cryptography Reference
In-Depth Information
1. Image generation A cryptographic hash visualization function h maps
an input x of arbitrary length to an output image y = h ( x ) of fixed size.
2. Efficiency Given h and an input x, h ( x ) is easy to compute.
3. Collision resistance For any output y , it is computationally infeasible to
find the input x such that h ( x ) ≈ y .
Additional methods for exchanging hashes over insecure channels
include a scheme based on the comparison of glyphs and also Phil Zim-
merman's biometric word lists . 49 Developed for the PGP free encryption
software, it is meant “to convey binary information in an authenticated
manner over a voice channel, such as a telephone, via biometric signatures.
The human voice that speaks the words, if recognized by the listener, serves
as a means of biometric authentication of the data carried by the words.” 50
The system is patterned along the lines of the alphabet used in the military
to unambiguously transmit letters over noisy channels. The PGP biometric
word list contains 256 words chosen for their linguistic distance, each
representing the 256 bytes values of 0 to 255. A public key fingerprint of,
for example, hexadecimal values E582: 94F2: E9A2: 2748: 6E8B:: 061B:
31CC: 528F: D7FA: 3F19 would be spoken as “topmost Istanbul: Pluto
vagabond: treadmill Pacific: brackish dictator: goldfish Medusa:: afflict
bravado: chatter revolver: Dupont midsummer: stopwatch whimsical:
cowbell bottomless.”
Visual Cryptography
Introduced in 1994 by Moni Naor and Adi Shamir, visual cryptography
offers a new type of cryptographic scheme whereby the encrypted mate-
rial is “decoded directly by the human visual system.” 51 The system con-
ceals the plaintext picture by splitting it into two distinct layers: a
ciphertext, printed on paper, and a secret key, printed on a transparency.
Taken individually, each layer looks like random noise, but overlaying
them one over the other reveals the plaintext (see figure 7.5). Like Ver-
nam's one-time pad, the system is perfectly secure in the information
theoretic sense. That is, even after having intercepted the ciphertext, an
attacker can do no better than flip a coin to guess the value of each
pixel of the plaintext.
The shares are constructed by substituting each pixel of the plaintext
with blocks of four subpixels, following the rules in figure 7.6: for each
Search WWH ::




Custom Search