[go: up one dir, main page]
More Web Proxy on the site http://driver.im/ Skip to main content

Advertisement

Springer Nature Link
Log in

A bi-directional associative memory based multiple image watermarking on cover video

  • Published:
Multimedia Tools and Applications Aims and scope Submit manuscript

Abstract

Telemedicine is a rapidly developing application of clinical medicine where medical information is transferred through the internet and other networks for the purpose of consulting, and remote medical procedures or examinations. This paper presents a novel neural network inspired watermarking technique, to enhance the authentication of the transmitted sensitive medical images over telemedicine network. In order to examine the medical images, the multiple scan images of same type or different types can be considered as watermarks and the same are trained using Bidirectional Associative Memory (BAM) neural network. The resultant weight matrix is embedded in the less correlated Principal Component Analysis (PCA) components of the wavelet coefficients of the cover video frames using additive embedding type of watermark. At the receiver end, the same BAM neural network is used with the randomly generated target matrix and the extracted weight matrix as input. As a result of this, the input matrix will be obtained. From the input matrix, the physician can extract their own information or scan images using their private or secret key. The proposed watermarking technique is validated with the existing systems in terms of imperceptibility, robustness and watermark capacity using the metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Normalized Cross Correlation (NCC), Bit Correction Rate (BCR), Detection rate, Receiver Operating Characteristics (ROC) and payload. The performance of the proposed system is evaluated by introducing various notable image processing attacks, geometrical attacks and video processing attacks on watermarked video. The experimental results demonstrate that the proposed technique has good perceptual quality of 50.6292 dB in terms of PSNR value, robustness of about nearly 1.0000 in terms of NCC value and payload of 1000 watermarks.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
£29.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price includes VAT (United Kingdom)

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14
Fig. 15
Fig. 16
Fig. 17
Fig. 18
Fig. 19
Fig. 20
Fig. 21

Similar content being viewed by others

References

  1. Agilandeeswari L, Ganesan K, Muralibabu K (2013) A side view based video in video watermarking using dwt and hilbert transform. Int Conf Secur Comput Commun Commun Computer Inf Sci, Springer Series 366–377

  2. Agilandeeswari L, Muralibabu K (2013) A robust video watermarking algorithm for content authentication using Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD). Int J Secur Its Appl 7(4):145–158

    Google Scholar 

  3. Bender W, Gruhl D, Morimoto N, Lu A (1996) Techniques for data hiding. IBM Syst J 35(3/4):313–337

    Article  Google Scholar 

  4. Boreczky JS, Rowe LA (1996) Comparison of video shot boundary detection techniques. J Electron Imaging 5(2):122–128

    Article  Google Scholar 

  5. Carnec M, Le Callet P, Barba D (2008) Objective quality assessment of color images based on a generic perceptual reduced reference. Signal Process Image Commun 23(4):239–256

    Article  Google Scholar 

  6. Carosi M, Pankajakshan V, Autrusseau F (2010) Towards a simplified perceptual quality metric for watermarking applications. Proc SPIE Conf Electron Imaging 7542:1–9

    Google Scholar 

  7. Chen Y, Chen J (2010) A novel blind watermarking scheme based on neural networks for image. IEEE Int Conf Inf Theory Inf Secur 548–552

  8. Chuan-Yu C, Sheng-Jyun S (2005) The application of a full counter propagation neural network to image while preserving high diagnostic quality. Proc Netw Sens Control 993–998

  9. Cox J, Kilian J, Leighton T, Shamoon T (1997) Secure spread spectrum watermarking for multimedia. IEEE Trans Image Process 6:1673–1687

    Article  Google Scholar 

  10. Cox IJ, Miller ML, Bloom (2002) Digital watermarking. Morgan Kaufmann Publisher, San Francisco

    Google Scholar 

  11. Dhavale SV, Deodhar RS, Pradhan D, Patnaik LM (2014) High payload adaptive audio watermarking based on cepstral feature modification. J Inf Hiding Multimed Sig Process 5(4):586–602

    Google Scholar 

  12. Dumitrescu S, Xiaolin W, Wang Z (2003) Detection of LSB steganography via sample pair analysis. IEEE Trans Sig Process 51(7):1995–2007

    Article  MATH  Google Scholar 

  13. El ‘Arbi M, Ben Amar C, Nicolas H (2006) Video watermarking based on neural networks. Multimed EXPO 1577–1580

  14. Fan Z, Hongbin Z (2005) Applications of a neural network to watermarking capacity of digital image. Neurocomputing 67:345–349

    Article  Google Scholar 

  15. Garg V, Brewer J (2011) Telemedicine security: a systematic review. J Diabetes Sci Technol 5(3):768–778

    Article  Google Scholar 

  16. Guo J-M, Prsateyo H (2014) False-positive-free SVD-based image watermarking. J Vis Commun Image Represent 25:1149–1163

    Article  Google Scholar 

  17. Huang HC, Chang FC (2013) Hierarchy-based reversible data hiding. Expert Syst Appl 40(1):34–43

    Article  Google Scholar 

  18. Huang HC, Fang WC (2011) Authenticity preservation with histogram-based reversible data hiding and quadtree concepts. Sensors 11(10):9717–9731

    Article  Google Scholar 

  19. Huang J, Shi YQ (2002) Reliable information bit hiding. IEEE Trans Circuits Syst Video Technol 12(10):916–920

    Article  Google Scholar 

  20. Hung-Hsu T, Chi-Chih L (2011) Wavelet-based image watermarking with visibility range estimation based on HVS and neural network. Pattern Recogn 44:751–763

    Article  MATH  Google Scholar 

  21. Irany BM, Guo XC, Hatzinakos D (2011)A high capacity reversible multiple watermarking scheme for medical images. 17th Int Conf Digit Signal Process 1–6

  22. Jin C, Su T, Pan L (2007) Multiple digital watermarking scheme based on ICA. 8th International Workshop on Image Analysis for Multimedia Interactive Services, Santorini, Greece 70–73

  23. Kasturi R, Jain R (1991) “Dynamic Vision”, in computer vision: principles. IEEE Computer Society Press, Washington

    Google Scholar 

  24. Latif A (2013) An adaptive digital image watermarking scheme using fuzzy logic and tabu search. J Inf Hiding Multimed Sig Process 4(4):250–271

    Google Scholar 

  25. Lee YK, Chen LH (2000) High capacity image steganographic model. Vis Image Sig Process IEEE Proc 147:288–294

    Article  Google Scholar 

  26. Lei B, Tan E-L, Chen S, Ni D, Wanga T, Lei H (2014) Reversible watermarking scheme for medical image based on differential evolution. Expert Syst Appl 41:3178–3188

    Article  Google Scholar 

  27. Lin C-C, Chang C-C, Chen Y-H (2014) A novel SVD-based watermarking scheme for protecting rightful ownership of digital images. J Inf Hiding Multimed Sig Process 5(2):124–143

    Google Scholar 

  28. Lin CY, Wu M, Bloom J, Cox I, Miller M, Liu Y (2001) Rotation, scale, and translation resilient watermarking for images. IEEE Trans Image Process 10(5):767–782

    Article  MATH  Google Scholar 

  29. Lu C, Hsu C (2007) Near-optimal watermark estimation and its countermeasure: antidisclosure watermark for multiple watermark embedding. IEEE Trans Circuits Syst Video Technol 17(4):454–467

    Article  Google Scholar 

  30. Mohamed KALLEL, Mohamed Salim BOUHLEL, Jean-Christophe LAPAYRE (2010) Use of multi-watermarking schema to maintain awareness in a teleneurology diagnosis platform. Radio Eng 19(1):68–73

    Google Scholar 

  31. Nallagarla R, Varadarajan S (2012) The robust digital image watermarking scheme with back propagation neural network in DWT domain. Int Conf Model Optim Comput Procedia Eng 38:3769–3778

    Google Scholar 

  32. Ngo NM, Unoki M, Miyauchi R, Suzuki Y (2014) Data hiding scheme for amplitude modulation radio broadcasting systems. J Inf Hiding Multimed Sig Process 5(3):324–341

    Google Scholar 

  33. Petitcolas FAP, Anderson RJ, Kuhn MG (1999) Information hiding—a survey. Proc IEEE 87:1062–1077

    Article  Google Scholar 

  34. Quan L, Jiang X (2006) Design and realization of a meaningful digital watermarking algorithm based on RBF neural network. Proc Sixth World Congr Intell Control Autom 1:2878–2881

    Article  Google Scholar 

  35. Rupachandra Singha T, Manglem Singh K, Roya S (2013) Video watermarking scheme based on visual cryptography and scene change detection. Int J Electron Commun 67:645–651

    Article  Google Scholar 

  36. Santhi V, Arulmozhivarman P (2013) Hadamard transform based adaptive visible/invisible watermarking scheme for digital images. J Inf Secur Appl 18(4):167–179

    Google Scholar 

  37. Sivanandam SN, Paulraj M (2012) Introduction to neural networks and its applications, 3rd edn. Tata Mc Graw Hill, New Delhi

  38. Swanson MD, Zhu B, Tewfik AH (1998) Multiresolution scene-based video watermarking using perceptual models. IEEE J Sel Areas Commun 16(4):540–50

    Article  MATH  Google Scholar 

  39. Takahashi A, Nishimura R, Suzuki Y (2005) Multiple watermarks for stereo audio signals using phase-modulation techniques. IEEE Trans Sig Processi 53(2):806–815

    Article  MathSciNet  Google Scholar 

  40. Wang Z, Bovik AC, Sheikh HR, Simoncelli EP (2004) Image quality assessment: from error visibility to structural similarity. IEEE Trans Image Process 13(4):600–612

    Article  Google Scholar 

  41. Wang S, Cui C, Niu X (2014) Watermarking for DIBR 3D images based on SIFT feature points. Measurement 48:54–62

    Article  Google Scholar 

  42. Wang J, Lian S (2012) On the hybrid multi-watermarking. Sig Process 92:893–904

    Article  Google Scholar 

  43. Weng S, Pan J-S (2014) Reversible watermarking based on multiple prediction modes and adaptive watermark embedding. Multimed Tools Appl 72(3):3063–3083

    Article  Google Scholar 

  44. Wolfgang RB, Podilchuk CI, Delp EJ (1999) Perceptual watermarks for digital images and video. IEEE Proc 87(7):1108–26

    Article  Google Scholar 

  45. Xinhong Z, Fan Z (2005) A blind watermarking algorithm based on neural network. Proc Int Conf Neural Netw Brain 2:1073–1076

    Google Scholar 

  46. Yu P, Tsai H, Lin J (2001) Digital watermarking based on neural networks for color images. Sig Proc 81(3):663–671

    Article  MATH  Google Scholar 

  47. Zhang Y (2009) Blind watermark algorithm based on HVS and RBF neural network in DWT domain. WSEAS Trans Comput 8(1):493–496

    Google Scholar 

  48. Zhang F, Hongbin Z (2004) Applications of neural network to watermarking capacity. Proc IEEE Int Symp Communi Inf Technol 1:340–343

    Google Scholar 

  49. Zhang L, Yan X, Li H, Chen M (2012) A dynamic multiple watermarking algorithm based on DWT and HVS. Int J Commun Netw Syst Sci 5(8):490–495

    Google Scholar 

  50. Zhang F, Zhang X (2007) Performance evaluation of multiple watermarks system. 2nd Workshop on Digital Media and its Application in Museum and Heritage, Chongqing, China 15–18

  51. Zhenfei W, Nengchao W, Baochang S (2006) A novel blind watermarking scheme based on neural network in wavelet domain. Proc 2006 Sixth World Congr Intell Control Autom 1:3024–3027

    Article  Google Scholar 

  52. Zhi-Ming Z, Rong-Yan L, Lei W (2003) Adaptive watermark scheme with RBF neural networks. Proc Int Conf Neural Netw Sig Process 2:1517–1520

    Google Scholar 

Download references

Acknowledgments

The authors would like to thank TIFAC-CORE in Automotive Infotronics located at VIT University, Vellore, 632014, India for providing necessary hardware and software facilities to carry out this work successfully.

Author information

Authors and Affiliations

  1. School of Information Technology and Engineering (SITE), VIT University, Vellore, 632014, Tamil Nadu, India

    L. Agilandeeswari

  2. TIFAC-CORE in Automotive Infotronics, VIT University, Vellore, 632014, Tamil Nadu, India

    K. Ganesan

Authors
  1. L. Agilandeeswari
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. K. Ganesan
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to L. Agilandeeswari.

Appendices

Appendix A. Shot boundary detection

1.1 Algorithm: shot boundary detection (motion frames = scenechangedetection (cv))

The statistical difference based scene change detection [4, 23] method is used here and it is given in detail as below:

Appendix B. Performance evaluation metrics

2.1 Imperceptibility analysis

The metrics used to measure Imperceptibility are Peak Signal-to-Noise Ratio (PSNR) [2, 6] and Structural Similarity Index (SSIM) [5, 16, 26, 36, 41] where the first measure is a subjective one, while the later one is an objective measure, whose value varies between 0 (No match) to 1 (Exact Match). PSNR is the commonly used metric to evaluate the degradation caused by various attacks. From the literature it is clear that, the minimum acceptable PSNR value is 20 dB [40].

$$ PSNR=10{ \log}_{10}\left(\frac{255^{\wedge }2}{MSE}\right) $$
(B.1)

where, Mean Square Error (MSE) between the cover video frame cvf(t) and attacked watermarked video frame awvf(t) is defined as,

$$ MSE=\frac{1}{T}\left({\displaystyle \sum_{t=1}^T{\left(cvf(t)- awvf(t)\right)}^2}\right) $$
(B.2)

where, T is total number of pixels in each frame.

The SSIM of two images x and y is given as,

$$ SSIM\left(x,y\right)=\frac{\left(2{L}_x{L}_y\right)\left(2{V}_{xy}\right)}{\left({L}_x^2+{L}_y^2\right)\left({V}_x^2+{V}_y^2\right)} $$
(B.3)

  • where, Lx and Ly represents the luminance factor of two images i.?e., the mean of x and y

  • Vx and Vy represents the contrast factor of two images i.?e., the standard deviation of x and y

  • Vxy represents the correlation coefficient between x and y

2.2 Robustness analysis

The metrics used to measure the robustness of the algorithm against possible attacks are Normalized Cross Correlation (NCC) and Bit Correction Rate (BCR). Both the metric’s acceptable value varies between 0 and 1. For the both NCC and BCR metric, value 1 indicates the identical or correlated image and value 0 represents that the two images are not identical or uncorrelated.

The Normalized Cross Correlation can be computed using the following equation as,

$$ NCC=\frac{{\displaystyle \sum \left(\left(O{W}_i-O{W}_m\right)\left(E{W}_i-E{W}_m\right)\right)}}{\sqrt{{\displaystyle \sum {\left(O{W}_i-O{W}_m\right)}^2}}\sqrt{{\displaystyle \sum {\left(E{W}_i-E{W}_m\right)}^2}}} $$
(B.4)

where, OW i is the intensity of the ith pixel in the image 1 (original watermark), EW i is the intensity of the ith pixel in image 2 (extracted watermark), OW m is the mean intensity of image 1 (original watermark), and EW m is the mean intensity of image 2 (extracted watermark).

Similarly, the Bit Correction Rate (BCR) can be computed using the equation as,

$$ BCR\left(OW,EW\right)=1-\frac{{\displaystyle \sum_{i=1}^m\left|O{W}_i-E{W}_i\right|}}{m} $$
(B.5)

where, OW i is the intensity of the ith pixel in image 1 (original watermark), EW i is the intensity of the ith pixel in image 2 (extracted watermark) and m is the size of an image.

Appendix C. Watermarking capacity

The embedding capacity of the proposed approach is evaluated based on the size of the weight matrix created, which in turn depends on the number of input and output neurons of BAM network.

It is possible to embed watermark of size 1000*1000 but it will reduce the number of watermarks to be embedded. That is, when there is an increase in size of image to be embedded, it will also increase the size of resultant weight matrix, thus the decrease in the number of watermarks. For our video of size 1024*1024, when we apply 2-level DWT, we can embed all 1000 images of size 256*256, but when the size increases to 512*512, only 50 watermarks can be embedded. Then, for an image size of about 1000*1000, we can embed only 2 watermarks. The results are tabulated in Table. C.1. The above said criteria can also produce good payload value, when we are either reducing the level of DWT or increasing the size of the cover video.

This can be expressed using the below equation:

$$ S=\frac{{\mathrm{W}}_{\mathrm{s}}}{N} $$
(C.1)

where,

S:

watermark size

N:

number of watermarks

Ws :

size of Weight matrix

Table. C.1. Evaluation of payload

For cover video of size 1024*1024

Maximum bits = 256* 256* 24 = 15, 72, 864 bits

DWT level

Watermark size

Total number of watermarks

Number of input neurons

Number of output neurons

Size of the weight matrix

Number of bits

Suitable for embedding in the present cover video

2-level DWT

256*256

10

65536

4

65536*4

262144

Number of bits < max bits can embed

30

65536

5

65536*5

327680

Number of bits < max bits can embed

50

65536

6

65536*6

393216

Number of bits < max bits can embed

100

65536

7

65536*7

458752

Number of bits < max bits can embed

200

65536

8

65536*8

524288

Number of bits < max bits can embed

500

65536

9

65536*9

589824

Number of bits < max bits can embed

1000

65536

10

65536*10

655360

Number of bits < max bits can embed

512*512

10

262144

4

262144*4

1048576

Number of bits < max bits can embed

30

262144

5

262144*5

1310720

Number of bits < max bits can embed

50

262144

6

262144*6

1572864

<=max bits can embed

1000*1000

1

1000000

1

1000000*1

1000000

Number of bits < max bits can embed

2

1000000

1

1000000*1

1000000

Number of bits < max bits can embed

1-level DWT

512*512

Maximum bits = 512* 512* 24 = 62, 91, 456 bits

     

10

262144

4

262144*4

1048576

Number of bits < max bits can embed

30

262144

5

262144*5

1310720

Number of bits < max bits can embed

50

262144

6

262144*6

1572864

Number of bits < max bits can embed

100

262144

7

262144*7

1835008

Number of bits < max bits can embed

200

262144

8

262144*8

2097152

Number of bits < max bits can embed

500

262144

9

262144*9

2359296

Number of bits < max bits can embed

1000

262144

10

262144*10

2621440

Number of bits < max bits can embed

1000*1000

1

1000000

1

1000000*1

1000000

Number of bits < max bits can embed

2

1000000

1

1000000*1

1000000

Number of bits < max bits can embed

Appendix D. Rotation attack

Let us consider the object matrix be \( \left[\begin{array}{ccc}\hfill X\hfill & \hfill Y\hfill & \hfill 1\hfill \end{array}\right] \) and rotation matrix be \( \left[\begin{array}{ccc}\hfill cos\theta\ \hfill & \hfill sin\theta \hfill & \hfill 0\hfill \\ {}\hfill - sin\ \theta \hfill & \hfill cos\theta \hfill & \hfill 0\hfill \\ {}\hfill 0\hfill & \hfill 0\hfill & \hfill 1\hfill \end{array}\right] \). Then the resultant rotated object can be obtained as

$$ \left[\begin{array}{ccc}\hfill X^{\prime}\hfill & \hfill Y^{\prime}\hfill & \hfill 1\hfill \end{array}\right]=\left[\begin{array}{ccc}\hfill X\hfill & \hfill Y\hfill & \hfill 1\hfill \end{array}\right]\left[\begin{array}{ccc}\hfill cos\theta\ \hfill & \hfill sin\theta \hfill & \hfill 0\hfill \\ {}\hfill - sin\ \theta \hfill & \hfill cos\theta \hfill & \hfill 0\hfill \\ {}\hfill 0\hfill & \hfill 0\hfill & \hfill 1\hfill \end{array}\right] $$
(D.1)

If we want to rotate an object by an angle θ in clockwise direction we take θ as a positive integer and negative for anti-clockwise rotation. In our experiment to test the performance of the proposed algorithm against rotation attack we have used imrotate () function to rotate an image then we have computed the Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR). For better understanding, we have computed the percentage of black pixel formation due to the rotation attack and the results are shown in Table. D.1. From the table, our inference is that, when the percentage of black pixel is more, the Mean Square Error (MSE) will increase which leads to decrease in PSNR and SSIM. This is applicable for the increased rotation angles from 1, 2, 5, 10, 20, 30, 40, 50, 60, 80, 100, 110, 120, 130, 140 and 160°. On the otherhand, when we rotate an image by an angle of 90 and 180°, the MSE value is reduced so that the PSNR and SSIM values are increasing in this case. This is because when an image is rotated by an angle of 90 and 180°, the percentage of black pixels is nearly 0, thus the modified pixels is less and hence the Mean Square Error (MSE) is also less that leads to an increase in the value of PSNR and SSIM.

Consider original video frame as,

Table. D.1. Rotation attack

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Agilandeeswari, L., Ganesan, K. A bi-directional associative memory based multiple image watermarking on cover video. Multimed Tools Appl 75, 7211–7256 (2016). https://doi.org/10.1007/s11042-015-2642-1

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11042-015-2642-1

Keywords

Search

Navigation