**3. Experimental results**

Several greyscale images of size 512512 were used as host images. A quarter of the host image *Lena* was used as the test data. To provide a variety of embedding rate, the value of the control parameter is not fixed. Simulations generated by the proposed FBBE algorithm were first shown in the following subsection. Subsequently, a high-performance hiding scheme was examined.

### **3.1 Simulations of the FBBE algorithm**

Figure 5 depicts the relationship between peak signal-to-noise ratio (PSNR) and robustness parameter that generated by the proposed FBBE algorithm. The size of the block was 44. The figure indicated that the optimal PSNR value of 57.45 dB is achieved with =1.

Fig. 5. The relationship between PSNR and .

An overflow/underflow can be occurred during bit embedding if a pixel value of the host image is a little either less than 255 or larger than 0. To overcome the overflow/underflow issues, a pixel-shifting approach can be performed in the spatial domain before data

Several greyscale images of size 512512 were used as host images. A quarter of the host image *Lena* was used as the test data. To provide a variety of embedding rate, the value of

were first shown in the following subsection. Subsequently, a high-performance hiding

Figure 5 depicts the relationship between peak signal-to-noise ratio (PSNR) and robustness

<sup>0</sup> <sup>2</sup> <sup>4</sup> <sup>6</sup> <sup>8</sup> <sup>10</sup> <sup>12</sup> <sup>14</sup> <sup>30</sup>

Beta

.

The figure indicated that the optimal PSNR value of 57.45 dB is achieved with

that generated by the proposed FBBE algorithm. The size of the block was 44.

1 or subtracting from

is not fixed. Simulations generated by the proposed FBBE algorithm

Lena Goldhill Zelda Elaine Tank

1 and <sup>2</sup> are

<sup>2</sup> . Both

<sup>1</sup> or , *p* <sup>2</sup> *p*

> =1.

embedment. Namely, if a pixel value *p* in a host image satisfied either *p*

**2.2.3 Overflow/underflow issues** 

can be adjusted to a new value by adding to

**3.1 Simulations of the FBBE algorithm** 

two predetermined threshold values.

**3. Experimental results** 

the control parameter

scheme was examined.

35

Fig. 5. The relationship between PSNR and

40

45

PSNR (dB)

50

55

60

parameter

The PSNR value is approximately linear decreased as increased. Actually, the larger the value of , the more robust performance can be obtained by the proposed method. The PSNR is defined by

$$PSNR = 10 \times \log\_{10} \frac{2\,\text{S}\text{S}^2}{MSE},\tag{7}$$

where (ˆ .)),(),( <sup>1</sup> 1 1 <sup>2</sup> *<sup>N</sup> i M j jixjix MN MSE* Here *x i j*),( and ˆ *jix* ),( denote the pixel values of the

original image and the marked image. Figure 6 shows the marked images generated by the proposed method with =12. Their average PSNR value was 33.35 dB with an embedding rate of 0.125 bits per pixel (bpp). It can be seen that the perceptual quality was acceptable.

Fig. 6. The marked images generated by the proposed FBBE algorithm. (a) Lena, (b) Goldhill, (c) Zelda, (d) Elaine, and (e) Tank.

For comparison, two graceful schemes, namely, Ni et al.'s algorithm (Ni et al., 2008) and Zeng et al.'s approach (Zeng et al., 2010) are compared with our method. Table 1 indicates the performance comparison of these methods on three test images. From Fig. 5 and Table 1 we can see that the proposed method with =5 (or of which value being less than 6) provides the largest payload among these methods while the PSNR for the proposed method is superior to that for the other two techniques. Moreover, Table 1 shows that the average hiding capacity provided by the proposed method is two times that achieved by Zeng et al.'s approach (Zeng et al., 2010), and five times larger than that achieved by Ni et al.'s algorithm (Ni et al., 2008).

Robust Lossless Data Hiding by Feature-Based Bit Embedding Algorithm 549

**Watermarks Attacks Survived** 

Brightness (+90%) BCR = 87.45%

Brightness (-100%) BCR = 89.65%

Contrast (40%) BCR = 87.48%

Contrast (-15%) BCR = 78.18%

Posterized (8-level) BCR = 85.26%

 Interleaved (Odd) BCR = 54.14%

Equalized BCR = 80.78% **watermarks** 

**Attacks Survived** 

Cropping (50%) BCR =87.88 %

JPEG2000 (CR\*=8.33)

BCR=71.89%

JPEG (CR=5.54) BCR=75.36%

Uniform noise (5%) BCR = 78.94%

Gaussian noise (4%)

BCR = 74.38%

Edge sharpening BCR = 98.92%

Mean filtering (3×3) BCR = 98.34%


Table 1. Hiding performance (Payload/ PSNR) comparison between various methods.

To demonstrate the robustness performance of the proposed method, examples of extracted watermarks after various manipulations of the image are given in Table 2. A logo of size 6363 with 8 bits/pixel 2 colours was used as the test watermark, as shown in Fig. 7. The bit correct ratio (BCR) is also included. The BCR is defined by

$$BCR = \left(\stackrel{ab-1}{\sum} \overline{w\_l \oplus \widetilde{w}\_l} \bigwedge\_{a \times b} \right) \times 100\%,\tag{8}$$

Fig. 7. The test watermark.

where *wi* and *wi* <sup>~</sup> represent the values of the original watermark and the extracted watermark respectively, as well as the size of a watermark is *ba* . Note that a majority-vote decision was employed during bits extraction. Although the BCR for those watermarks, which extracted from the images that gone through attacks such as JPEG2000, JPEG, equalized, interleaved, and inversion are not high, they are identifiable. Although the BCR for the watermark extracted from an image which manipulated by inversion attack is only 1.99%, it is recognizable. Furthermore, Fig. 8 shows the BCR performance of the survived watermarks under a variety of degree of Uniform/Gaussian noise additions attacks. From the figure we can see that the proposed method is more robust against Uniform than Gaussian noise additions attacks. Similarly, Fig. 9 indicates the proposed method has the better performance in resisting JPEG200 than JPEG attacks. Figure 10 shows that the proposed method is nearly free from brightness attacks. Finally, Fig. 11 indicates that the extracted watermarks are tolerant of colour quantization attack even if the number of level of pixel-value in a marked image is reduced to 8.

4,480/ 40.47

16,384/ 38.09

32,768/ 41.56

*Lena Zelda Goldhill* Average

6,336/ 40.18

16,384/ 38.10

32,768/ 42.84

%,100

 

<sup>~</sup> represent the values of the original watermark and the extracted

5,717/ 40.28

16,384/ 38.09

32,768/ 42.04

(8)

Methods Images

6,336/ 40.19

16,384/ 38.07

32,768/ 41.71

> 

 

*BCR*

*ab i*

correct ratio (BCR) is also included. The BCR is defined by

Table 1. Hiding performance (Payload/ PSNR) comparison between various methods.

To demonstrate the robustness performance of the proposed method, examples of extracted watermarks after various manipulations of the image are given in Table 2. A logo of size 6363 with 8 bits/pixel 2 colours was used as the test watermark, as shown in Fig. 7. The bit

~ <sup>1</sup>

*ii*

 *ba ww*

watermark respectively, as well as the size of a watermark is *ba* . Note that a majority-vote decision was employed during bits extraction. Although the BCR for those watermarks, which extracted from the images that gone through attacks such as JPEG2000, JPEG, equalized, interleaved, and inversion are not high, they are identifiable. Although the BCR for the watermark extracted from an image which manipulated by inversion attack is only 1.99%, it is recognizable. Furthermore, Fig. 8 shows the BCR performance of the survived watermarks under a variety of degree of Uniform/Gaussian noise additions attacks. From the figure we can see that the proposed method is more robust against Uniform than Gaussian noise additions attacks. Similarly, Fig. 9 indicates the proposed method has the better performance in resisting JPEG200 than JPEG attacks. Figure 10 shows that the proposed method is nearly free from brightness attacks. Finally, Fig. 11 indicates that the extracted watermarks are tolerant of colour quantization attack even if the number of level

0

Ni et al.'s algorithm

Zeng et al.'s approach

Proposed Method

Fig. 7. The test watermark.

of pixel-value in a marked image is reduced to 8.

where *wi* and *wi*


Robust Lossless Data Hiding by Feature-Based Bit Embedding Algorithm 551

JPEG2000 attacks JPEG attacks

<sup>2</sup> <sup>3</sup> <sup>4</sup> <sup>5</sup> <sup>6</sup> <sup>7</sup> <sup>8</sup> <sup>9</sup> <sup>70</sup>

Fig. 9. The BCR for the proposed method under JPEG2000/JPEG attacks, respectively.

Compression ratio


The percentage of (attack) variation

Fig. 10. The BCR for the proposed method under Brightness attacks.

75

88

90

92

94

BCR (%)

96

98

100

80

85

BCR (%)

90

95

100


\* CR stands for compression ratio, which is defined as the ratio of the size of a host image to that of a compressed image.

The last four bits of the pixel in the marked image were truncated.

Table 2. Examples of watermarks extracted from image *Lena*. (=12)

Fig. 8. The BCR for the proposed method under Uniform/Gaussian noise additions attacks, respectively.

CR stands for compression ratio, which is defined as the ratio of the size of a host image to that of a

<sup>1</sup> 1.5 <sup>2</sup> 2.5 <sup>3</sup> 3.5 <sup>4</sup> 4.5 <sup>5</sup> <sup>70</sup>

Fig. 8. The BCR for the proposed method under Uniform/Gaussian noise additions attacks,

Noise additions (increament by %)

The last four bits of the pixel in the marked image were truncated.

Table 2. Examples of watermarks extracted from image *Lena*. (=12)

**Watermarks Attacks Survived** 

Inversion BCR = 1.99%

Interleaved (Even) BCR = 53.87%

**watermarks** 

Uniform noise attack Gaussizn noise attack

**Attacks Survived** 

Median filtering (3×3)

BCR = 98.76%

Quantization BCR = 95.67%

compressed image.

75

respectively.

80

85

BCR (%)

90

95

100

\*

Fig. 9. The BCR for the proposed method under JPEG2000/JPEG attacks, respectively.

Fig. 10. The BCR for the proposed method under Brightness attacks.

Robust Lossless Data Hiding by Feature-Based Bit Embedding Algorithm 553

Lena Goldhill Zelda Elaine Tank

Lena Goldhill Zelda Elaine Tank

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Payload (bpp)

0 5 10 15

Beta

.

Fig. 12. The trade-off between payload and PSNR for the proposed scheme.

30

0.1

Fig. 13. The relationship between payload and

0.2

0.3

0.4

0.5

Payload (bpp)

0.6

0.7

0.8

35

40

45

PSNR (dB)

50

55

60

Fig. 11. The BCR for the proposed method under (color) quantization attacks.

#### **3.2 Simulations of high-performance hiding scheme**

The trade-off between PSNR and payload for the proposed scheme was depicted in Figure 12. The figure indicated that the average PSNR achieved by the proposed scheme was approximately 55 dB at a bit rate of 0.236 bpp. Whereas, the optimal PSNR value of 37.76 dB can be achieved in image *Zelda* with bit rate of 0.747 bpp. In addition, the relationship between payload (or embedding rate) and robustness parameter was drawn in Fig. 13. From the figure we can see that the larger the value of , the higher the bit rate was achieved.

For comparison, three outstanding approaches: Wu et al.'s scheme (Wu et al. 2009), Lee et al.'s algorithm (Lee et al., 2010), and Yang & Tsai's technique (Yang & Tsai, 2010) were compared with our method. Performance comparison between these methods was given in Table 3. It is obvious that the proposed method provides the largest payload among these methods while the PSNR for the proposed method is superior to that for the other three algorithms. Moreover, Table 3 implies that the hiding capacity provided by the proposed method is approximately two times that achieved by the Wu *et al.*'s scheme (Wu et al. 2009), and is two times that achieved by Lee et al.'s algorithm (Lee et al., 2010). Moreover, Table 4 revealed the superiority of our scheme when the PSNR value around 43 dB. The average embedding rate for the proposed scheme was two times larger than that for the Wu et al.'s technique (Wu et al. 2009).

<sup>5</sup> <sup>10</sup> <sup>15</sup> <sup>20</sup> <sup>25</sup> <sup>30</sup> <sup>35</sup> <sup>40</sup> <sup>45</sup> <sup>50</sup> <sup>85</sup>

The trade-off between PSNR and payload for the proposed scheme was depicted in Figure 12. The figure indicated that the average PSNR achieved by the proposed scheme was approximately 55 dB at a bit rate of 0.236 bpp. Whereas, the optimal PSNR value of 37.76 dB can be achieved in image *Zelda* with bit rate of 0.747 bpp. In addition, the relationship

For comparison, three outstanding approaches: Wu et al.'s scheme (Wu et al. 2009), Lee et al.'s algorithm (Lee et al., 2010), and Yang & Tsai's technique (Yang & Tsai, 2010) were compared with our method. Performance comparison between these methods was given in Table 3. It is obvious that the proposed method provides the largest payload among these methods while the PSNR for the proposed method is superior to that for the other three algorithms. Moreover, Table 3 implies that the hiding capacity provided by the proposed method is approximately two times that achieved by the Wu *et al.*'s scheme (Wu et al. 2009), and is two times that achieved by Lee et al.'s algorithm (Lee et al., 2010). Moreover, Table 4 revealed the superiority of our scheme when the PSNR value around 43 dB. The average embedding rate for the proposed scheme was two times larger than that for the Wu et al.'s

was drawn in Fig. 13.

, the higher the bit rate was

Fig. 11. The BCR for the proposed method under (color) quantization attacks.

**3.2 Simulations of high-performance hiding scheme** 

between payload (or embedding rate) and robustness parameter

From the figure we can see that the larger the value of

90

achieved.

technique (Wu et al. 2009).

95

100

Fig. 12. The trade-off between payload and PSNR for the proposed scheme.

Fig. 13. The relationship between payload and .

Robust Lossless Data Hiding by Feature-Based Bit Embedding Algorithm 555

brightness/contrast, mean/median filtering, and inversion. Furthermore, the payload and PSNR provided by the proposed two methods outperform those provided by existing

The proposed two methods can be extended to color images by embedding data bits in the RGB system separately. In addition, to further enlarge the hiding storage of the FBBE algorithm, an extra one (or two) data bits could be hidden in each IWT coefficients block during data embedment. However, a tradeoff between PSNR and payload size may be a problem with this algorithm. These issues will be discussed in detail in future work. Furthermore, to reduce memory space and transmission delay, the decreasing of the

Alattar, A. M. (2004). Reversible watermark using the difference expansion of a generalized integer transform. *IEEE T. Image Processing*, Vol. 13, No. 8, pp. 1147-1156. Al-Qaheri, H.; Mustafi, A. & Banerjee, S. (2010). Digital Watermarking using Ant Colony

Calderbank, A.R.; Daubechies, I.; Sweldens, W. & Yeo, B.L. (1998). Wavelet transforms that

Cox, I.J.; Miller, M.L.; Bloom, J.A.; Fridrich, J. & Kalker T. (Ed(s.)) (2008). *Digital* 

Fan, L.; Gao, T. & Yang Q. (2011). A novel watermarking scheme for copyright protection

Hu, Y., Lee; H. K. & Li, J. (2009). DE-based reversible data hiding with improved overflow

Hsiao, J. Y.; Chan, K.F. & Chang, J.M. (2009). Block-based reversible data embedding. *Signal* 

Lai, C.C.; Huang, H.C. & Tsai, C.C. (2010). A digital watermarking scheme based on singular

Lee, C.F.; Chen, H.L. & Tso, H.K. (2010). Embedding capacity raising in reversible data

Lin, C.C. & Shiu, P.F. (2010). High capacity data hiding scheme for DCT-based images.

Liu, J. C. & Shih, M. H. (2008). Generalization of pixel-value differencing staganography for data hiding in images. *Fundamenta Informaticate*, Vol. 83, pp. 319-335. Martinez-Noriega, R.; Nakano, M.; Kurkoski, B. & Yamaguchi, K. (2011). High Payload

*Watermarking and Steganography*, 2nd Ed., Morgan Kaufmann., MA.

scheme. *ICIC Express Letters*, Vol. 3, No. 3 (A), pp. 397-402.

*Computing Information and Control*, Vol. 5, No. 7, pp. 1867-1873.

*Multimedia Signal Processing*, Vol. 1, No. 3, pp. 179-189.

Optimization in Fractional Fourier Domain. *Journal of Information Hiding and* 

map integers to integers. *Applied & Computational Harmonics Analysis*, Vol. 5, No. 3,

based on adaptive joint image feature and visual secret sharing. *International Journal of Innovative Computing, Information and Control,* Vol. 7, No. 7(A), pp. 3679-3694. Gu, Q. & Gao, T. (2009). A novel reversible watermarking algorithm based on wavelet lifting

location map. *IEEE T. Circuits and Systems for Video Technology*. Vol. 19, No. 2, pp.

value decomposition and micro-genetic algorithm. *International Journal of Innovative* 

hiding based on prediction of different expansion. *The Journal of Systems and* 

*Journal of Information Hiding and Multimedia Signal Processing*, Vol. 1, No. 3, pp. 220-

Audio Watermarking: toward Channel Characterization of MP3 Compression. *Journal of Information Hiding and Multimedia Signal Processing*, Vol. 2, No. 2, pp. 91-

schemes.

**5. References** 

overhead bits will be our future study.

pp.332-369.

250-260.

240.

107.

*Processing*, Vol. 89, pp. 556-569.

*Software*, Vol. 83, pp. 1864-1872.


Table 3. Embedding rate and PSNR performance comparison between various methods when PSNR value was approximately 48 dB.


Table 4. Embedding rate and PSNR performance comparison between various methods when PSNR value was approximately 43 dB.
