Page 173 - 2024-Vol20-Issue2
P. 173
169 | Muttashar & Fyath
Fig. 5. Brief description of the fusion process.
Fig. 4. Examples of the 9D-chaotic attractor for the C1 -C2 iv. Finally, a random scrambling technique is applied to the
plane. image ICF to produce the encrypted digital image IDEnc. Note
that the scrambler arranges the array pixels randomly without
operation with RGB chaotic image, and scrambling. The pro- altering their values.
cedures for these steps are as follows
3) Optical Encryption Subscheme
i. A fusion process is applied to the input double-color im- The optical encryption subscheme is based on double chaotic
age, yielding two new color images IF1 and IF2. This method phase encoding (DCPE) implemented in the 2D optical Fourier
is adopted according to the amount of information carried by transform (FT) domain for each RGB channel of the received
various binary bits of image pixels and reflects scrambling digital encrypted image, as shown in Fig. 6. Note that the 2D
at the bit-slice levels. Fig. 5 shows the process of the fusion optical FT can be implemented using a single lens. The lens
method used in this work. produces an FT image at the back focal plane for an image
located at the front focal plane. For each color channel, the
ii. The image combiner gathers the two-color fusion images optical encryption (OE) consists of two cascaded FTs sup-
IF1 and IF2, each of size M×N, one above the other, to produce ported by two chaotic phase masks (CPMs). Each CPM is
a single fusion image IF with a size of 2M×N. generated by one output of the chaotic system. One CMP is
bonded with the primary image, and another is placed in the
iii. A chaotic color image IC of size, 2M×N is constructed as Fourier domain, as shown in Fig. 7.
an encryption key by three of the nine chaotic system outputs The mathematician framework describes the operation of
(C1, C2 and C3, where each output is responsible for one of DCPE-based FT encryption, which can be adopted from af-
the RGB channels. The chaotic image IC is XORing with ter replacing the DRPE with double chaotic phase masks
the color fusion image IF to get a chaotic fusion image ICF . (DCPMs). Consider the red channel as an example; the CPMs
The chaotic gray images ICR, ICG and ICB, corresponding to can be explained as follow [27]
the RGB channels, are constructed based on the sequence
C1, C2 and C3, respectively. This is done after using integer CPM1 R(x, y) = exp( jf1(x, y)) = exp( j2pC4(x, y)), (3)
sequencing according to (2)
CPM2 R(x, y) = exp( jf2(x, y)) = exp( j2pC5(x, y)), (4)
ICR(i) = f ix(mod(C1(i)1015, 255) (2)
ICG(i) = f ix(mod(C2(i)1015, 255) where C4 and C5 are the fourth and fifth chaotic. The same
ICB(i) = f ix(mod(C3(i)1015, 255), operations are applied to green and blue channels. The first
step is to bound the input image whose electric field Ein with
where f ix(C) rounds each element of C to the nearest integer CPM1, the combined function is FT, leading to An electric
and mod(a, m) returns the remainder after dividing a by m. field E1(u, v)
Here, a and m represent the dividend and the divisor, respec-
tively. E1(u, v) = FT [Ein(x, y)CPM1 R(x, y)]
= [Ein(x, y)epx(2p jCCPM1 R(x, y)) (5)
exp(-2p j(ux + uv))]dxdy