Page 17 - 2024-Vol20-Issue2
P. 17
13 | Yousif, Hameed & Al-Zuhairi
TABLE V.
RESULTS OF ENCRYPTION/DECRYPTION TIME MEASURED FOR TEST IMAGES WITH DIFFERENT SIZES
Image Size Encryption time (Second) Decryption time (Second)
Cameraman 256x256 0.154946 0.670657
Lena 512x512 0.527656 1.686384
Baboon 1024x1024 1.580442 2.502797
Barbara 256x256 0.162288 0.643419
Boat 512x512 0.378125 1.590626
Peppers 1024x1024 1.466533 2.325564
256x256 0.131102 0.697361
512x512 0.384833 1.627816
1024x1024 1.453171 2.238646
256x256 0.132878 0.728806
512x512 0.390394 1.66202
1024x1024 1.426099 2.303133
256x256 0.149949 0.672086
512x512 0.339793 1.670934
1024x1024 1.508857 2.39036
256x256 0.143553 0.650492
512x512 0.356593 1.628605
1024x1024 1.462144 2.261963
TABLE VI. information entropy, correlation coefficient in horizontal di-
RESULTS OF PSNR AND CORRELATION BETWEEN INPUT rection, MSE, PSNR, UACI, and NPCR. The results of these
metrics are computed for Lena image of size 256 × 256. It
AND RECONSTRUCTED IMAGES UNDER CROPPING can be noticed from the outcomes in this table that the entropy
ATTACK. score after the zigzag scan stage remains constant (7.4318
in Table II, while this value increases after the second and
Size of Cropping PSNR (dB)) Correlation third encryption stages to: 7.9950 and 7.9971, respectively.
Cropping ½ 10.0189 0.3957 Also, the correlation value decreases from 0.9492 (Table III
12.9694 0.6001 to: 0.1330, 0.0204 and 4.9266 × 10-4, respectively after the
Cropping ¼ in bottom right 13.1048 0.6026 first, second and third ciphering stages. Further, the value of
Cropping ¼ in top left 13.1152 0.6241 MSE after the zigzag scan is 4.5449 × 103 . Then, this value
Cropping ¼ in center 18.9409 0.8577 increases to: 7.7540 × 103 and 1.3962 × 104, respectively
Cropping 1/8 in top left after applying the RSA and chaotic systems. Contrariwise,
the PSNR score decreases from 11.5556 at the first layer
and 0.8577, respectively for PSNR and correlation. Also, (zigzag scan) to: 9.2356 and 6.6813, respectively after em-
different outcomes are obtained when changing the cropping ploying the other encryption phases. Finally, for the UACI
location. For instance, cropping 1/4 from the bottom right of and NPCR, their values increase gradually from 27.4992 and
the cipher image yields 12.9694 dB and 0.6001 for PSNR and 99.37, respectively to: 33.3258 and 100, respectively after
correlation, respectively, cropping 1/4 from the top left yields implementing the RSA and chaotic maps.
13.1048 dB and 0.6026 for PSNR and correlation, respectively,
while cropping 1/4 from the center part yields 13.1152 dB and From Table VII, it can be concluded that each technique
0.6241 for PSNR and correlation, respectively. Hence, the has a significant role in the encryption process. The modified
reconstructed image conserves the significant data included zigzag scan does not effect on the entropy value, but it is
in the original image even if the cutting size reaches to half obviously effect on the other metrics. The impact of RSA
size of the cipher image thereby the TIC endures efficiently mechanism on the increasing the entropy, MSE, UACI, NPCR
the cropping attack. values, and decreasing the correlation and PSNR values is
extremely clear. Finally, the high dimension chaotic maps
F. Case Study of the Proposed TIC Method with confusion/diffusion processes can further improve the
Table VII is presented in order to show the effect of each
encryption stage of the proposed TIC algorithm in terms of