Edgy Photos Statistical Steganalysis 1
Alaa A. Jabbar, 2Shahrin Bin Sahib, 3Mazdak Zamani Faculty of Information & Communication Technology, Universiti Teknikal Malaysia Melaka, 76100 Melaka, Malaysia,
[email protected],
[email protected] 3 Advanced Informatics School, Universiti Teknologi Malaysia, Malaysia,
[email protected] 1,2
Abstract In order to analyze how the correlativity behavioral pattern would change after data embedding, Edgy theme photos have been embedded from 1 to 5 bits out of eight bits per channel.
Keywords: artificial intelligence, multimedia security, steganography, watermarking 1. Introduction The known themes of photos are solid, smooth, and edgy photos. Furthermore, results of pixel’s correlation analysis shows that there is difference between photos with wide smooth area and photos with partial smooth area. Thus the four analyzed photo themes for extracting patterns are wide flat region, wide smooth region, partial smooth region, and edgy photos [1-5]. This paper elaborates details of a proposed steganalysis method and then studies characteristics of the photo themes. Layer is the main concept for correlativity analysis in offered method. All the pixels located in the same layer of neighbors have same importance and if only one of the pixels in the layer be correlated to the central pixel then that layer and all outer layers would be counted as correlated. To approve correlativity of a layer at least one of the pixels in that layer or one of the pixels in inner layers should be correlated to the central pixel [6-13]. Pixel color matching or in another word AND color analysis examines the correlativity of the central pixel with neighbors in different layers in a way that all red, green, and blue channels of both pixels should be the same. This analysis starts from comparing the central pixel with first neighbor pixel in first layer. Then the process continues with comparing color channels of the central pixel with next neighbors in the same layer and then the neighbors in next layers up to fifth layer [14-23]. Channel color matching, or in other hand OR correlativity analysis, measures correlativity of the central pixel with neighbor pixels up to five layers in a way that if red, green, or blue channel values of the central pixel be the same as one of neighbors then that layer up to fifth layer would be considered as OR correlated layers [24-31]. The correlativity analysis results of 50 samples of 1-bit embedded Edgy photos from dataset are shown in Table 1. In addition, Figure 1 illustrates how the results are scattered for each layer of AND correlativity analysis after 1-bit secret data embedding. Figure 2 illustrates OR correlativity analysis results of Edgy dataset photos after 1-bit secret data embedding. Since normally there are more true results for OR comparison than AND comparison then in Figure 2 representative marks of OR comparison results are more condensed than AND. Table 2 presents the required statistical analysis results calculated based on Table 1 values for calculating correlativity behavioral pattern of 1-bit embedded Edgy photos. The pattern is presented in Table 3. Table 1. Correlativity analysis results of 1-bit embedded Edgy photos File No 1 2 3 4 5 6 7 8 9
Total Analyzed Pixels 4827630 5941130 5941130 2272900 2272900 2043700 4054500 747650 1736800
Neighbor-1
Neighbor-2
Neighbor-3
AND
OR
AND
OR
AND
OR
5.20 8.86 21.30 3.06 7.76 18.11 0.05 6.66 4.33
62.19 87.16 97.34 43.51 66.53 88.38 48.55 60.95 52.36
8.31 14.80 34.20 4.57 13.06 29.03 0.11 10.26 7.76
80.64 94.58 99.30 71.15 83.75 95.53 80.98 79.74 73.85
10.64 19.56 42.43 5.23 16.90 36.03 0.18 12.82 10.56
89.56 96.98 99.72 85.96 90.95 98.01 93.79 88.12 86.56
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
1027848 4054500 2272900 2043700 2272900 1892100 1892100 1736800 6057740 5576350 2272900 768612 1272700 1003300 638820 768612 2272900 2272900 1892100 4054500 2272900 4054500 4054500 4054500 2043700 2272900 2272900 4054500 2043700 4054500 4054500 2272900 1003300 1003300 1206500 768612 768612 768612 768612 1027848 2087500
2.88 5.60 4.09 2.66 8.79 11.70 6.32 10.98 10.32 0.96 3.77 4.14 1.81 7.59 1.59 7.67 3.82 1.13 11.09 26.76 9.18 6.39 7.92 13.02 14.19 0.94 5.92 12.61 5.49 7.19 14.58 7.43 7.62 4.15 13.60 12.01 1.57 9.01 2.32 1.81 9.45
58.25 79.76 57.35 74.27 67.78 83.95 72.89 92.76 69.84 48.78 57.96 50.78 46.60 70.81 46.79 61.24 73.06 44.01 74.20 95.51 76.75 70.64 91.01 88.21 78.89 42.44 78.34 96.31 71.41 70.40 83.43 83.85 69.07 57.19 85.04 78.51 29.08 64.24 39.61 42.81 68.55
4.22 9.17 7.16 4.62 14.23 17.72 9.88 18.04 16.02 1.51 5.86 6.73 2.40 11.24 2.20 10.84 7.08 1.82 18.81 41.30 14.80 9.88 12.33 19.99 21.75 1.39 9.45 19.94 10.34 13.03 23.66 10.26 12.02 7.09 20.51 22.58 2.52 16.40 3.66 2.79 14.94
83.41 92.46 79.06 92.04 84.46 93.15 88.94 98.17 84.98 74.66 80.05 76.10 71.82 86.61 71.28 80.19 91.09 73.65 88.97 98.64 90.35 86.48 98.30 95.96 89.74 65.47 91.06 99.37 92.57 87.82 92.68 95.04 84.96 80.91 94.22 92.30 51.23 82.46 62.16 69.46 86.81
5.33 11.95 9.76 6.73 18.35 21.12 12.34 23.01 19.07 1.94 7.39 8.77 2.90 13.49 2.68 12.55 10.07 2.58 24.42 48.97 18.90 12.17 15.90 24.64 26.22 1.70 12.15 25.25 14.67 17.52 29.34 12.09 15.38 9.79 24.56 32.14 3.33 23.02 4.84 3.62 18.49
Figure 1. AND correlativity analysis of 1-bit embedded Edgy photos
93.56 96.42 89.18 97.07 92.00 96.56 94.75 99.25 91.95 88.09 90.48 88.04 85.19 93.19 84.55 89.78 96.40 90.12 94.69 99.41 95.24 93.12 99.55 98.10 94.70 79.25 95.47 99.78 97.44 93.79 96.37 98.15 92.08 90.92 97.13 96.52 68.56 90.88 76.97 83.33 94.20
Figure 2. OR correlativity analysis of 1-bit embedded Edgy photos Table 2. Analysis statistics of 1-bit embedded Edgy photos Neighbor layer 1 Neighbor layer 2 Neighbor layer 3 AND OR AND OR AND OR 7.508 67.9868 12.0456 84.772 15.2698 92.2376 Average 26.76 97.34 41.3 99.37 48.97 99.78 Maximum 0.05 29.08 0.11 51.23 0.18 68.56 Minimum 5.43014 16.93563 8.618321 10.51543 10.60951 6.2512 Standard deviation Table 3. Extracted correlativity behavioural pattern of 1-bit embedded Edgy photos Min to left SD Left to right SD Right SD to max AND 0.05 ≤ X < 2.07786 2.07786 ≤ X ≤ 12.93814 12.93814 < X ≤ 26.76 Neighbor layer 1 OR 29.08 ≤ X < 51.05117 51.05117 ≤ X ≤ 84.92243 84.92243 < X ≤ 97.34 AND 0.11 ≤ X < 3.427279 3.427279 ≤ X ≤ 20.66392 20.66392 < X ≤ 41.3 Neighbor layer 2 OR 51.23 ≤ X < 74.25657 74.25657 ≤ X ≤ 95.28743 95.28743 < X ≤ 99.37 AND 0.18 ≤ X < 4.660295 4.660295 ≤ X ≤ 25.87931 25.87931 < X ≤ 48.97 Neighbor layer 3 OR 68.56 ≤ X < 85.9864 85.9864 ≤ X ≤ 98.4888 98.4888 < X ≤ 99.78 Table 4 contains correlativity analysis results of Edgy photos after 2-bits secret data embedding. Scattering of AND correlativity evaluation values of Table 4 are shown in Figure 3 and OR correlativity values in Figure 4. Figure 3 shows that the AND comparison results are layered and more separated than OR comparison results in Figure 4. The required statistical values for generating the pattern after 2-bits embedding, based on Table 4, are presented in Table 5 and calculated pattern in Table 6. Table 4. Correlativity analysis results of 2-bits embedded Edgy photos File No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Total Analyzed Pixels 4827630 5941130 5941130 2272900 2272900 2043700 4054500 747650 1736800 1027848 4054500 2272900 2043700 2272900 1892100 1892100
Neighbor-1
Neighbor-2
Neighbor-3
AND
OR
AND
OR
AND
OR
1.36 2.67 5.19 0.73 1.86 4.04 0.05 1.47 1.03 0.78 1.84 1.11 0.94 2.07 2.86 1.69
64.26 87.19 96.90 44.24 68.36 88.94 48.39 62.03 52.86 60.49 80.23 60.41 73.79 68.61 83.44 74.16
2.62 5.45 11.21 1.64 4.04 8.64 0.09 2.97 2.24 1.39 3.61 2.26 1.97 4.57 5.87 3.24
82.29 95.11 99.39 71.61 84.87 96.12 80.85 80.36 74.10 84.44 92.83 81.14 92.01 84.88 93.28 89.66
3.80 8.28 17.17 2.59 6.39 13.18 0.15 4.52 3.57 1.99 5.32 3.47 3.25 7.27 8.71 4.65
90.53 97.31 99.76 86.22 91.49 98.28 93.72 88.45 86.64 93.95 96.60 90.39 97.13 92.22 96.64 95.06
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
1736800 6057740 5576350 2272900 768612 1272700 1003300 638820 768612 2272900 2272900 1892100 4054500 2272900 4054500 4054500 4054500 2043700 2272900 2272900 4054500 2043700 4054500 4054500 2272900 1003300 1003300 1206500 768612 768612 768612 768612 1027848 2087500
2.97 2.39 0.39 0.94 1.08 0.47 1.77 0.50 1.69 1.15 0.34 2.61 6.36 2.30 1.86 2.59 3.63 3.17 0.33 1.77 3.94 1.55 2.02 3.69 2.00 1.91 1.03 3.02 2.73 0.44 2.13 0.63 0.48 2.20
92.20 71.16 49.82 58.64 53.30 47.74 72.20 47.25 61.88 72.70 45.70 75.43 95.27 78.91 71.87 90.95 88.15 79.84 42.94 78.90 95.69 73.55 71.54 83.73 84.77 70.18 60.17 85.34 81.88 30.11 67.62 41.28 43.69 70.06
6.22 5.18 0.68 1.85 2.21 0.79 3.59 0.86 3.14 2.66 0.65 5.95 13.98 4.72 3.55 4.97 7.40 6.49 0.58 3.49 8.24 3.76 4.47 7.66 3.55 3.96 2.24 6.47 6.01 0.82 4.58 1.21 0.96 4.73
98.31 85.91 75.39 80.48 77.57 72.30 87.48 71.64 80.72 91.28 74.86 90.04 98.75 91.57 87.31 98.47 96.13 90.58 65.84 91.52 99.36 93.22 88.33 93.09 95.56 85.89 82.54 94.69 94.01 52.28 84.84 63.34 69.81 87.71
9.42 7.98 0.92 2.76 3.39 1.12 5.20 1.20 4.47 4.41 1.07 9.63 21.66 7.13 5.19 7.39 11.20 9.47 0.81 5.16 12.73 6.31 7.00 11.54 4.81 6.03 3.57 9.84 9.85 1.24 7.36 1.80 1.48 7.26
99.32 92.42 88.45 90.67 88.73 85.48 93.58 84.78 90.10 96.51 90.65 95.24 99.45 95.82 93.54 99.62 98.18 95.17 79.48 95.67 99.78 97.67 94.01 96.57 98.36 92.58 91.66 97.36 97.28 69.22 92.19 77.64 83.55 94.60
Figure 3. AND correlativity analysis of 2-bits embedded Edgy photos
Figure 4. OR correlativity analysis of 2-bits embedded Edgy photos Table 5. Analysis statistics of 2-bits embedded Edgy photos
Average Maximum Minimum Standard deviation
Neighbor layer 1 AND OR 1.9154 68.9752 6.36 96.9 0.05 30.11 1.296528 16.55992
Neighbor layer 2 AND OR 3.9886 85.4752 13.98 99.39 0.09 52.28 2.814635 10.3657
Neighbor layer 3 AND OR 6.0942 92.5944 21.66 99.78 0.15 69.22 4.342097 6.152464
Table 6. Extracted correlativity behavioural patterns of 2-bits embedded Edgy photos Min to left SD Left to right SD Right SD to max AND 0.05 ≤ X < 0.618872 0.618872 ≤ X ≤ 3.211928 3.211928 < X ≤ 6.36 Neighbor layer 1 OR 30.11 ≤ X < 52.41528 52.41528 ≤ X ≤ 85.53512 85.53512 < X ≤ 96.9 AND 0.09 ≤ X < 1.173965 1.173965 ≤ X ≤ 6.803235 6.803235 < X ≤ 13.98 Neighbor layer 2 OR 52.28 ≤ X < 75.1095 75.1095 ≤ X ≤ 95.8409 95.8409 < X ≤ 99.39 AND 0.15 ≤ X < 1.752103 1.752103 ≤ X ≤ 10.4363 10.4363 < X ≤ 21.66 Neighbor layer 3 OR 69.22 ≤ X < 86.44194 86.44194 ≤ X ≤ 98.74686 98.74686 < X ≤ 99.78 Table 7 holds correlativity analysis results of Edgy photos after embedding 3-bits of secret data in each byte of cover photo. Values of this table are used in Figure 5 and Figure 6 to depict how the results are scattered. The figures show that 3-bits embedding will dramatically change the correlativity in comparison with less embedded or original copy of cover photos. The results are much more layered and for AND analysis condensed in lower part of the figure while for OR analysis in upper part. Table 8 contains the statistical required results calculated from values of Table 7 to produce the correlativity behavioral pattern of 3-bits embedded Edgy photos in Table 9. Table 7. Correlativity analysis results of 3-bits embedded Edgy photos File No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Total Analyzed Pixels 4827630 5941130 5941130 2272900 2272900 2043700 4054500 747650 1736800 1027848 4054500 2272900 2043700 2272900 1892100 1892100 1736800 6057740 5576350 2272900 768612 1272700 1003300 638820 768612 2272900 2272900 1892100 4054500 2272900 4054500 4054500 4054500
Neighbor-1
Neighbor-2
Neighbor-3
AND
OR
AND
OR
AND
OR
0.34 0.63 0.97 0.17 0.40 0.79 0.02 0.32 0.25 0.23 0.45 0.29 0.28 0.46 0.60 0.43 0.70 0.49 0.14 0.24 0.31 0.13 0.43 0.18 0.42 0.33 0.11 0.55 1.03 0.52 0.42 0.62 0.73
63.19 82.80 91.67 44.25 66.90 84.93 47.34 60.24 51.54 59.94 76.81 60.22 70.82 66.27 79.27 72.04 87.23 69.39 49.91 57.62 53.51 47.37 70.09 46.71 59.87 69.80 46.16 73.04 90.66 76.75 69.38 86.27 84.24
0.73 1.51 2.45 0.40 0.95 1.91 0.06 0.77 0.61 0.48 1.12 0.67 0.67 1.14 1.39 0.96 1.67 1.16 0.27 0.53 0.68 0.28 0.96 0.35 0.83 0.86 0.25 1.39 2.89 1.19 1.00 1.56 1.91
83.01 95.14 99.32 71.99 85.36 96.23 80.43 80.47 74.21 84.68 92.91 82.20 91.74 84.99 93.17 89.90 98.16 86.41 75.82 80.61 78.33 72.44 87.75 71.81 80.74 91.16 75.45 90.33 98.70 92.08 87.48 98.39 96.14
1.16 2.55 4.26 0.69 1.62 3.25 0.10 1.26 1.02 0.77 1.77 1.11 1.20 2.03 2.33 1.53 2.80 1.98 0.40 0.86 1.13 0.45 1.57 0.52 1.24 1.57 0.47 2.51 5.04 2.00 1.58 2.59 3.19
91.12 97.49 99.78 86.45 91.92 98.42 93.57 88.63 86.74 94.15 96.75 91.11 97.19 92.36 96.68 95.29 99.35 92.80 88.75 90.81 89.28 85.56 93.86 84.99 90.21 96.65 91.02 95.53 99.47 96.21 93.76 99.67 98.26
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
2043700 2272900 2272900 4054500 2043700 4054500 4054500 2272900 1003300 1003300 1206500 768612 768612 768612 768612 1027848 2087500
0.64 0.12 0.48 0.81 0.41 0.44 0.69 0.52 0.44 0.28 0.64 0.66 0.14 0.53 0.20 0.14 0.46
76.67 42.90 75.81 90.57 73.07 69.38 80.06 81.57 67.54 59.90 81.48 79.42 30.49 66.40 41.15 43.45 68.29
1.49 0.23 1.13 2.16 1.14 1.22 1.83 1.11 1.02 0.68 1.47 1.49 0.28 1.24 0.41 0.33 1.15
90.84 66.18 91.64 99.34 93.66 88.50 93.24 95.62 86.02 83.30 94.68 94.57 52.93 85.69 63.91 69.94 87.99
2.43 0.35 1.86 3.67 2.08 2.05 3.04 1.66 1.73 1.18 2.47 2.54 0.47 2.14 0.65 0.55 1.97
95.40 79.74 95.86 99.79 97.90 94.19 96.74 98.45 92.82 92.19 97.49 97.68 69.75 92.85 78.03 83.66 94.80
Figure 5. AND correlativity analysis of 3-bits embedded Edgy photos 100 80 60
Neighbors in 1 Layer Neighbors in 2 Layers
40
Neighbors in 3 Layers 20 0 0
10
20
30
40
50
Figure 6. OR correlativity analysis of 3-bits embedded Edgy photos Table 8. Analysis statistics of 3-bits embedded Edgy photos Neighbor layer 1 Neighbor layer 2 Neighbor layer 3 AND OR AND OR AND OR 0.4316 66.8876 1.0396 85.712 1.7478 92.8234 Average 1.03 91.67 2.89 99.34 5.04 99.79 Maximum 0.02 30.49 0.06 52.93 0.1 69.75 Minimum 0.228673 15.05971 0.606465 10.23873 1.048338 6.081768 Standard deviation Table 9. Extracted correlativity behavioral patterns of 3-bits embedded Edgy photos
Neighbor layer 1 Neighbor layer 2 Neighbor layer 3
AND OR AND OR AND OR
Min to left SD 0.02 ≤ X < 0.202927 30.49 ≤ X < 51.82789 0.06 ≤ X < 0.433135 52.93 ≤ X < 75.47327 0.1 ≤ X < 0.699462 69.75 ≤ X < 86.74163
Left to right SD 0.202927 ≤ X ≤ 0.660273 51.82789 ≤ X ≤ 81.94731 0.433135 ≤ X ≤ 1.646065 75.47327 ≤ X ≤ 95.95073 0.699462 ≤ X ≤ 2.796138 86.74163 ≤ X ≤ 98.90517
Right SD to max 0.660273 < X ≤ 1.03 81.94731 < X ≤ 91.67 1.646065 < X ≤ 2.89 95.95073 < X ≤ 99.34 2.796138 < X ≤ 5.04 98.90517 < X ≤ 99.79
Table 10 reflects the correlativity analysis values of 4-bits secret data embedded Edgy photos. Figure 7 and Figure 8 visually show where the analysis results of Table 10 are located. The AND correlativity values are more condensed and almost more overlapped than 3-bits embedded results. Also there is big difference between distance of layer-1/layer-2 and layer-2/layer-3. Table 11 holds statistical required values calculated from Table 10 needed for generating the correlativity behavioral pattern of 4-bits embedded Edgy photos. The pattern is presented in Table 12. Table 10. Correlativity analysis results of 4-bits embedded Edgy photos File No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
Total Analyzed Pixels 4827630 5941130 5941130 2272900 2272900 2043700 4054500 747650 1736800 1027848 4054500 2272900 2043700 2272900 1892100 1892100 1736800 6057740 5576350 2272900 768612 1272700 1003300 638820 768612 2272900 2272900 1892100 4054500 2272900 4054500 4054500 4054500 2043700 2272900 2272900 4054500 2043700 4054500 4054500 2272900 1003300 1003300 1206500
Neighbor-1
Neighbor-2
Neighbor-3
AND
OR
AND
OR
AND
OR
0.07 0.12 0.15 0.04 0.07 0.12 0.03 0.06 0.05 0.06 0.15 0.06 0.07 0.08 0.11 0.08 0.13 0.09 0.04 0.05 0.09 0.04 0.15 0.06 0.10 0.07 0.03 0.10 0.23 0.10 0.13 0.18 0.20 0.11 0.03 0.10 0.21 0.09 0.13 0.18 0.10 0.13 0.10 0.17
56.16 69.19 75.17 42.02 58.37 70.99 44.94 53.10 46.35 54.51 65.93 54.35 61.67 57.17 66.73 62.70 72.27 60.50 47.12 51.84 49.87 43.93 61.49 43.76 53.06 61.02 44.22 63.16 75.18 65.76 60.74 72.09 70.89 64.94 40.74 64.65 74.57 64.71 60.47 67.99 69.08 58.82 54.82 68.94
0.18 0.32 0.40 0.12 0.23 0.38 0.06 0.17 0.13 0.13 0.34 0.19 0.21 0.25 0.28 0.21 0.33 0.23 0.09 0.15 0.21 0.09 0.31 0.12 0.21 0.25 0.09 0.26 0.58 0.28 0.28 0.42 0.46 0.33 0.08 0.29 0.50 0.30 0.32 0.43 0.30 0.28 0.21 0.38
81.81 93.09 97.64 71.75 84.03 94.63 78.62 79.14 73.34 83.43 91.08 81.38 89.68 83.62 91.32 88.41 96.15 85.32 75.06 79.36 78.02 71.52 86.52 70.96 79.11 89.69 75.11 89.06 97.23 90.76 85.94 96.37 94.63 89.36 65.87 90.01 97.49 92.62 86.99 91.74 93.75 84.44 82.63 92.88
0.29 0.59 0.75 0.19 0.40 0.65 0.09 0.31 0.25 0.25 0.54 0.30 0.38 0.44 0.49 0.38 0.59 0.41 0.14 0.24 0.36 0.15 0.50 0.19 0.32 0.45 0.17 0.49 1.02 0.46 0.45 0.73 0.77 0.54 0.13 0.49 0.86 0.54 0.55 0.72 0.46 0.48 0.38 0.64
91.19 97.30 99.72 86.59 92.00 98.39 92.89 88.57 86.70 94.16 96.70 91.37 96.97 92.32 96.58 95.32 99.23 92.95 88.72 90.71 89.63 85.46 93.91 84.88 89.87 96.72 91.22 95.65 99.45 96.36 93.69 99.56 98.30 95.37 79.85 95.80 99.75 98.05 94.15 96.72 98.30 92.68 92.57 97.40
45 46 47 48 49 50
768612 768612 768612 768612 1027848 2087500
0.16 0.04 0.12 0.07 0.03 0.09
67.92 29.60 58.58 38.88 40.55 59.84
0.38 0.10 0.29 0.13 0.08 0.23
93.16 53.05 84.89 63.46 69.10 86.65
0.64 0.16 0.49 0.22 0.16 0.42
97.89 70.21 93.30 78.32 83.63 94.75
Figure 7. AND correlativity analysis of 4-bits embedded Edgy photos
Figure 8. OR correlativity analysis of 4-bits embedded Edgy photos Table 11. Analysis statistics of 4-bits embedded Edgy photos Neighbor layer 1 Neighbor layer 2 Neighbor layer 3 AND OR AND OR AND OR 0.0994 58.427 0.2518 84.4374 0.4324 92.8364 Average 0.23 75.18 0.58 97.64 1.02 99.75 Maximum 0.03 29.6 0.06 53.05 0.09 70.21 Minimum 0.050281 10.91739 0.119552 9.808643 0.207331 6.016857 Standard deviation Table 12. Extracted correlativity behavioural patterns of 4-bits embedded Edgy photos Min to left SD Left to right SD Right SD to max AND 0.03 ≤ X < 0.049119 0.049119 ≤ X ≤ 0.149681 0.149681 < X ≤ 0.23 Neighbor layer 1 OR 29.6 ≤ X < 47.50961 47.50961 ≤ X ≤ 69.34439 69.34439 < X ≤ 75.18 AND 0.06 ≤ X < 0.132248 0.132248 ≤ X ≤ 0.371352 0.371352 < X ≤ 0.58 Neighbor layer 2 OR 53.05 ≤ X < 74.62876 74.62876 ≤ X ≤ 94.24604 94.24604 < X ≤ 97.64 AND 0.09 ≤ X < 0.225069 0.225069 ≤ X ≤ 0.639731 0.639731 < X ≤ 1.02 Neighbor layer 3 OR 70.21 ≤ X < 86.81954 86.81954 ≤ X ≤ 98.85326 98.85326 < X ≤ 99.75 Table 13 represents the correlativity analysis results of 5-bits embedded Edgy photos. Figures 5.9 and 5.10, which are depicted based on Table 13 values, show that AND analysis values are very condensed in lower part of the figure and completely overlapped while for OR analysis results of the
layers are more separated than before. Also like 4-bits embedded copies the distance between layer1/layer-2 is much more than the distance between layer-2/layer-3. Table 14 constituted from statistical values, calculated from Table 13, required to generate the correlativity behavioral pattern of 5-bits embedded Edgy photos. The pattern is presented in Table 15. Table 13. Correlativity analysis results of 5-bits embedded Edgy photos File No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
Total Analyzed Pixels 4827630 5941130 5941130 2272900 2272900 2043700 4054500 747650 1736800 1027848 4054500 2272900 2043700 2272900 1892100 1892100 1736800 6057740 5576350 2272900 768612 1272700 1003300 638820 768612 2272900 2272900 1892100 4054500 2272900 4054500 4054500 4054500 2043700 2272900 2272900 4054500 2043700 4054500 4054500 2272900 1003300 1003300 1206500 768612 768612 768612 768612 1027848 2087500
Neighbor-1
Neighbor-2
Neighbor-3
AND
OR
AND
OR
AND
OR
0.01 0.01 0.02 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.02 0.01 0.02 0.01 0.00 0.01
42.69 48.82 51.74 36.00 43.50 49.79 36.59 40.88 36.87 42.27 47.29 42.07 46.77 42.71 47.55 46.17 50.32 44.90 38.89 40.77 40.58 36.14 45.53 36.86 41.02 45.87 38.37 46.34 51.46 47.65 44.72 50.22 49.62 46.84 34.83 46.75 51.14 47.41 44.30 48.19 49.09 43.47 42.29 48.77 49.12 26.53 44.17 32.84 33.73 44.76
0.03 0.05 0.06 0.02 0.03 0.05 0.02 0.03 0.02 0.02 0.05 0.03 0.04 0.03 0.04 0.04 0.05 0.04 0.02 0.02 0.04 0.02 0.04 0.03 0.04 0.04 0.02 0.04 0.07 0.04 0.04 0.06 0.06 0.04 0.02 0.04 0.07 0.04 0.05 0.06 0.04 0.04 0.03 0.04 0.07 0.02 0.04 0.03 0.02 0.04
74.84 84.06 87.95 68.05 76.43 85.31 71.91 72.88 68.35 76.33 82.31 74.72 82.32 76.15 82.34 80.45 86.36 78.06 70.20 73.17 73.00 66.99 79.07 67.02 72.54 81.79 71.82 81.24 87.82 82.29 78.36 86.76 85.60 80.98 62.37 81.43 87.64 83.84 79.11 83.30 84.33 76.96 76.02 83.79 84.25 51.32 77.65 60.36 64.55 79.07
0.06 0.09 0.11 0.05 0.06 0.09 0.03 0.06 0.05 0.05 0.08 0.06 0.08 0.07 0.08 0.07 0.10 0.07 0.04 0.05 0.07 0.04 0.07 0.05 0.06 0.08 0.05 0.08 0.12 0.08 0.07 0.11 0.11 0.08 0.03 0.08 0.12 0.08 0.08 0.10 0.08 0.06 0.06 0.08 0.12 0.04 0.09 0.05 0.04 0.07
89.23 95.73 98.20 85.47 90.20 96.62 89.77 87.17 85.36 92.26 94.71 89.60 95.52 90.56 94.70 93.62 97.43 91.41 87.06 88.91 88.43 84.02 92.23 83.80 87.83 95.10 90.21 94.17 98.02 94.66 91.83 97.82 96.80 93.60 78.59 94.01 98.10 96.32 92.40 95.11 96.26 90.85 90.96 95.61 96.22 69.73 91.82 77.58 82.25 92.98
Figure 9. AND correlativity analysis of 5-bits embedded Edgy photos
Figure 10. OR correlativity analysis of 5-bits embedded Edgy photos Table 14. Analysis statistics of 5-bits embedded Edgy photos Neighbor layer 1 Neighbor layer 2 Neighbor layer 3 AND OR AND OR AND OR 0.0102 43.704 0.0384 77.2692 0.072 91.2168 Average 0.02 51.74 0.07 87.95 0.12 98.2 Maximum 0 26.53 0.02 51.32 0.03 69.73 Minimum 0.005147 5.579125 0.014194 7.847811 0.023299 5.807567 Standard deviation Table 15. Extracted correlativity behavioural patterns of 5-bits embedded Edgy photos Min to left SD Left to right SD Right SD to max AND 0 ≤ X < 0.005053 0.005053 ≤ X ≤ 0.015347 0.015347 < X ≤ 0.02 Neighbor layer 1 OR 26.53 ≤ X < 38.12487 38.12487 ≤ X ≤ 49.28313 49.28313 < X ≤ 51.74 AND 0.02 ≤ X < 0.024206 0.024206 ≤ X ≤ 0.052594 0.052594 < X ≤ 0.07 Neighbor layer 2 OR 51.32 ≤ X < 69.42139 69.42139 ≤ X ≤ 85.11701 85.11701 < X ≤ 87.95 AND 0.03 ≤ X < 0.048701 0.048701 ≤ X ≤ 0.095299 0.095299 < X ≤ 0.12 Neighbor layer 3 OR 69.73 ≤ X < 85.40923 85.40923 ≤ X ≤ 97.02437 97.02437 < X ≤ 98.2
2. Conclusion This paper aimed to extract the behavioral patterns of the four photo themes after secret data embedding. The behavioral patterns extracted from pixel correlativity analysis (AND-analysis) and color channel correlativity analysis (OR-analysis) results of the photo samples in dataset. The dataset is constituted from four categories of photo themes which there are six groups of photos in each. Original photos, 1-bit embedded, 2-bits embedded, 3-bits embedded, 4-bits embedded, and 5-bits embedded photos are the photo groups in this theme. Each theme contains 50 photo samples and the dataset altogether has 1200 samples which the twenty four behavioral patterns are extracted from them. In future work, the extracted behavioral patterns will be used in developing a steganalysis tool to examine the given photo objects for likelihood of carrying secret data and in next step for prediction of embedding ratio.
3. References [1] Mazdak Zamani, Azizah Bt Abdul Manaf, Rabiah Bt Ahmad, Farhang Jaryani, Saman Shojae Chaeikar Hossein Rouhani Zeidanloo. “Genetic Audio Watermarking”. Book: Information Processing and Management. Springer Berlin Heidelberg. 514-517. 2010. [2] Mazdak Zamani, Azizah Bt Abdul Manaf, Shahidan M. Abdullah, Saman Shojae Chaeikar. Mazdak Technique for PSNR Estimation in Audio Steganography. Applied Mechanics and Materials. Vols. 229-231 (2012). pp 2798-2803. Trans Tech Publications, Switzerland. [3] Alaa A. Jabbar, Shahrin Bin Sahib, Mazdak Zamani. Pixel Correlation Behavior in Different Themes. Communications in Computer and Information Science. Springer Berlin Heidelberg. ISSN: 1865-0929. 2013. (2013 International Conference on Geo-Informatics in Resource Management & Sustainable Ecosystem [GRMSE2013]. Nov 08-Nov 10, 2013. Wuhan, China.) [4] Alaa A. Jabbar, Shahrin Bin Sahib, Mazdak Zamani. Correlation Analysis of the Four Photo Themes in Five Layers. Communications in Computer and Information Science. Springer Berlin Heidelberg. ISSN: 1865-0929. 2013. (2013 International Conference on Geo-Informatics in Resource Management & Sustainable Ecosystem. Nov 08-Nov 10, 2013. Wuhan, China.) [5] Mazdak Zamani, Azizah Abdul Manaf, Rabiah Ahmad, Akram Zeki, and Shahidan Abdullah. A Genetic-Algorithm-Based Approach for Audio Steganography. World Academy of Science, Engineering and Technology. 30 2009. ISSN: 2070-3740. Pages: 355-358. 24-26 June 2009. [6] Mazdak Zamani, Hamed Taherdoost, Azizah A. Manaf, Rabiah B. Ahmad, and Akram M. Zeki. An Artificial-Intelligence-Based Approach for Audio Steganography. Journal of Open Problems in Science and Engineering. Volume: 1 Issue: 1. Oct 2009. ISSN: 2077-1118. Pg 64-68. [7] Mazdak Zamani, Azizah Abdul Manaf, Rabiah Ahmad, Farhang Jaryani, Hamed Taherdoost, Saman Shojae Chaeikar, Hossein Rouhani Zeidanloo. “A Novel Approach for Genetic Audio Watermarking”. Journal of Information Assurance and Security 5 (2010). Pages 102-111. [8] Mazdak Zamani, Azizah Abdul Manaf, Rabiah Ahmad, Farhang Jaryani, Hamed Taherdoost, Saman Shojae Chaeikar, Hossein Rouhani Zeidanloo. Genetic Audio Steganography. International Journal on Recent Trends in Engineering & Technology. Vol 3 (2), 2010. Pg 89-91. [9] Shahidan M. Abdullah, Azizah A. Manaf, Mazdak Zamani. Recursive Reversible Image Watermarking Using Enhancement of Difference Expansion Techniques. Journal of Information Security Research. Volume 1 Number 2. June 2010. Pg 64-70. ISSN: 0976-4143. [10] Mazdak Zamani, Azizah Abdul Manaf, Hossein Rouhani Zeidanloo, and Saman Shojae Chaeikar. “Genetic Subtitution-Based Audio Steganography for High-Capacity Applications”. International Journal for Internet Technology and Secured Transactions (IJITST). Pages 97-110. ISSN 1748 - 5703. Volume 3 Issue 1, April 2011. Inderscience Publishers, Geneva, Switzerland. [11] Akram M. Zeki, Azizah A. Manaf, Adamu A. Ibrahim and Mazdak Zamani. A Robust Watermark Embedding in Smooth Areas. Research Journal of Information Technology. Year: 2011. Volume: 3. Issue: 2. Page No.: 123-131. eISSN: 2151-7959. [12] Mazdak Zamani, Azizah Bt Abdul Manaf, Shahidan M. Abdullah. An Overview on Audio Steganography Techniques. International Journal of Digital Content Technology and its Applications. Advanced Institute of Convergence Information Technology. 2012.. [13] Alaa A. Jabbar, Shahrin Bin Sahib, Mazdak Zamani. Determining the threshold of pixel correlativity analysis. International Journal of Digital Content Technology and its Applications. Advanced Institute of Convergence Information Technology. Volume 7, Number 11, July 2013. [14] Mazdak Zamani; Azizah Abdul Manaf. "Genetic algorithm for fragile audio watermarking". Telecommunication Systems, Special Issue on Innovations in Emerging Multimedia Communication Systems, “In Press”. ISSN: 1018-4864. Publisher: Springer. [15] Mazdak Zamani, Azizah Abdul Manaf, and Rabiah Ahmad. “Knots of Substitution Techniques of Audio Steganography”. The 2009 International Conference on Telecom Technology and Applications. Pages 415-419. June 6-8, 2009. Manila, Philippines. [16] Mazdak Zamani, Azizah Abdul Manaf, and Rabiah Ahmad. “Current Problems of Substitution Technique of Audio Steganography”. International Conference on Artificial Intelligence and Pattern Recognition. ISBN: 978-1-60651-007-0. Pg 154-160. 13-16 July 2009. Florida, USA.
[17] Mazdak Zamani, Azizah Abdul Manaf, Rabiah Ahmad, Akram Zeki, and Shahidan Abdullah. “Genetic Algorithm as an Approach to Resolve the Problems of Substitution Techniques of Audio Steganography”. The 2009 International Conference on Genetic and Evolutionary Methods. 13-16 July 2009. Pages 170-175. Las Vegas, Nevada, USA. [18] Mazdak Zamani, Azizah Abdul Manaf, Rabiah Ahmad, Akram Zeki. An Approach to Improve the Robustness of Substitution Techniques of Audio Steganography. 2nd IEEE International Conference on Computer Science and Information Technology 2009. Vol 2. Pg 5-9. 8-11 Aug 2009. Beijing, China. [19] Mazdak Zamani, Hamed Taherdoost, Azizah Abdul Manaf, Rabiah Ahmad, and Akram Zeki. “Robust Audio Steganography via Genetic Algorithm”. Third International Conference on Information & Communication Technologies. Pages 149 - 153. 15-16 August 2009. Pakistan. [20] Mazdak Zamani, Azizah Abdul Manaf, Rabiah Ahmad, Akram Zeki, and Pritheega Magalingam. “A Novel Approach for Audio Watermarking”. Fifth International Conference on Information Assurance and Security. Pages 83-86. 18-20, August 2009. Xi'an, China. [21] Mazdak Zamani, Azizah Abdul Manaf, Rabiah Ahmad, Farhang Jaryani, Hamed Taherdoost, and Akram Zeki. “A Secure Audio Steganography Approach”. The 4th International Conference for Internet Technology and Secured Transactions. Pages 501-506. 9-12 Nov 2009. London, UK. [22] Akram M. Zeki, Azizah A. Manaf, and Mazdak Zamani. Bit-Plane Model: Theory and Implementation. Engineering Conference 2010 (EnCon2010). 14-16 April 2010. Malaysia. [23] Nuha Omran Abokhdair, Azizah Bt Abdul Manaf, Mazdak Zamani. Integration of Chaotic Map and Confusion Technique for Color Medical Image Encryption. 6th International Conference on Digital Content, Multimedia Technology and its Applications. 20-23. 16-18 Aug 2010. Korea. [24] Shahidan M. Abdullah, Azizah A. Manaf, and Mazdak Zamani. Capacity and Quality Improvement in Reversible Image Watermarking Approach. 6th International Conference on Networked Computing and Advanced Information Management. 81-85. 2010. Seoul, Korea. [25] Mazdak Zamani, Azizah Bt Abdul Manaf, Shahidan M. Abdullah, Saman Shojae Chaeikar. Correlation between PSNR and Bit per Sample Rate in Audio Steganography. 11th International Conference on Signal Processing. Pgs 163-168. France, April 2-4, 2012. [26] Mazdak Zamani, Azizah Bt Abdul Manaf, Shahidan M. Abdullah. Correlation between PSNR and Size Ratio in Audio Steganography. 11th International Conference on Telecommunications and Informatics. Pages 82-87. France, April 2-4, 2012. [27] Mazdak Zamani, Azizah Bt Abdul Manaf, Shahidan M. Abdullah. Efficient Embedding for Audio Steganography. 2nd International Conference on Environment, Economics, Energy, Devices, Systems, Communications, Computers, Mathematics. 195-199. France, April 2-4, 2012. [28] Mazdak Zamani, Azizah Abdul Manaf, and Rusni Daruis. “Azizah Technique for Efficiency Measurement in Steganography”. 8th International Conference on Information Science and Digital Content Technology 3, art. No. 6269320, pp. 480-484. June 26-28, 2012. Korea. [29] Alaa A. Jabbar, Shahrin Bin Sahib, Mazdak Zamani. An Introduction to Image Steganography Techniques. International Conference on Advanced Computer Science Applications and Technologies (ACSAT’12). 26-28 November 2012. [30] Alaa A. Jabbar, Shahrin Bin Sahib, Mazdak Zamani. Multimedia Data Hiding Evaluation Metrics. 7th WSEAS International Conference on Computer Engineering and Applications (CEA '13). Milan, Italy. January 9-11, 2013. [31] Alaa A. Jabbar, Shahrin Bin Sahib, Mazdak Zamani. An Introduction to Watermarking Techniques. 12th WSEAS International Conference on Applications of Computer Engineering (ACE '13). Cambridge, MA, USA. January 30 - February 1, 2013.
