Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
ß 2014 RPS
Received 29 November 2011; accepted 3 September 2013
DOI 10.1179/1743131X13Y.0000000062 The Imaging Science Journal 2014 VOL 62 NO 5 265
Jindal and Singh Joint image compression–encryption
3 Fractional order a vs PSNR of Lena image for different 5 Fractional order a vs PSNR of different images at CR
CRs using DFrFT of 50% using DFrFT
PSNR and ‘a’, for CRs of 10, 20, 30, 40, 50 and 75% yðm,nÞ~Fca3 fy2 ðm2 ,n2 ÞRðm2 ,n2 Þg (3)
respectively. Figures 42629 shows a plot between ‘a’ and
with y2 ðm2 ,n2 Þ~Fca2 fy1 ðm1 ,n1 ÞRðm1 ,n1 Þg
PSNR for different images at CRs of 10, 50 and 75%
and y1 ðm1 ,n1 Þ~Fca1 ff ðm0 ,n0 Þg
respectively.
Using the same fractional order at different CRs, the
images can be compressed. For CR of 2?5, 16 and 28% The resultant function yðm,nÞ can be regarded as the
and 2566256 image, the proposed algorithm has mean encrypted image.32 The decryption process is the reverse
square error (MSE) 0?0000, 0?02 and 0?063 respectively operation with respect to the encryption. First we take a
and Alfalou et al.3 (Fig. 7a) calculated 0?0289, 0?08 and DFrCT of order –a3 on the encrypted image yðm,nÞand
0?122 respectively. Therefore, an improvement in MSE multiplying the RPM R2 ðm,nÞ and then we get the
0?0289, 0?078 and 0?059 respectively was obtained. midterm function y2 ðm2 ,n2 Þ. Then we perform a DFrCT
of order –a2 on the function y2 ðm2 ,n2 Þ and multiplying
the RPM R1 ðm1 ,n1 Þ, thus function y1 ðm1 ,n1 Þ will be
DFrCT for encryption/decryption recovered. After another DFrCT of order –a1 on the
In electronic commerce, for exchange of information function y1 ðm1 ,n1 Þ, we finally get the original image
encryption is carried out for the security of data. In the (m0,n0). Here the random mask R2 ðm,nÞ and R1 ðm1 ,n1 Þ
proposed method, for encryption cascaded multistage are the complex conjugate of R1 ðm1 ,n1 Þ and R2 ðm2 ,n2 Þ
DFrCT with random phase filters have been utilised. A respectively. Figure 7 is the block diagram for image
primary image encryption technique31 involves the encryption and decryption using DFrCT.
process in which the primary image is encoded using The encryption and decryption are not only sensitive
three cascaded stages of fractional order with two to RPMs but also to the fractional orders. Encryption
intermediate randomly encoded phase masks. The n- and decryption results for Barbara, cameraman, Lena
stage of DFrCT can provide n-dimensional extra keys and rice images of 2566256 pixels using DFrCT are
indicated by the fractional orders. In case of 2D DFrCT obtained. Figure 8a is the original Lena image.
transform, there are two different fractional orders Encrypted form of this image is Fig. 8b. The encrypted
8 Simulation results of encryption for Lena image using DFrCT: a original Lena image; b encrypted Lena image;
c decrypted image with right key; d decrypted Image with wrong key
10 Simulation results of joint compression–encryption for Cameraman image using DFrFT and DFrCT respectively: a ori-
ginal Cameraman image; b compressed image (CR520%); c compressed–encrypted image; d decrypted–decompressed
image (with wrong key); e decrypted–decompressed image (with right key)
image Fig. 8b is messy in nature and shows high with right fractional orders. The graphical results show
encryption quality. The encryption quality is expressed that with change in the original fractional order, the
in terms of the total change in pixels values between the value of MSE increases in decryption. Decrypted image
plain-frame and cipher-frame. Figure 8c and d shows is sensitive to –a3 than to a1.
the simulation results of decrypted Lena image with The sensitivity of fractional order keys provides
right and wrong key respectively. security to proposed algorithm. However, compressed–
encrypted must be secured from various attacks. The
Joint image compression-encryption security in proposed algorithm is doubted in concern of
double random phase encoding. However, double
algorithm random phase encoding has been effectively used for
The proposed joint approach used for image compres- encryption34 and information hiding.35 Its weakness lies
sion–encryption has several advantages in image trans- in its linearity. The double random phase encoding
mission. Compression before encryption will remove potentially insecure the algorithm against known plain-
redundancy of data and security will increase. Figure 9 text attack and chosen plaintext attack and Brute force
describes the joint model of compression–encryption attack.
method using fractional transform. Brute force attack depends upon exhaustive keys
Numerical simulations have been performed on test search and is feasible only for the cryptosystems with
images Baboon, Peepers, baboon, cameraman and house to relatively small key space. The value of fractional part
examine the validity of this algorithm. The original image for each key can be increased upto any number. The
was the picture of Cameraman with 2566256 pixels, which increased key space size infeasible this attack. It has
is shown in Fig. 10a. Figure 10b shows the compressed been demonstrated elsewhere36 that Brute force attack
image with 20% CR and cut-off 1?23 with fractional order is also intractable in concern of double random phase
(o) 0?98. This compressed image is encrypted with fractional encoding.
keys {1, 0?93, 0?8} in Fig. 10c. The other fractional orders The inadequacy of proposed algorithm is its vulner-
can be used for compression and encryption. Figure 10d ability to known plaintext attack and chosen cipher text
and e is the decrypted–decompressed image with wrong key attack owing to double random phase encoding.36 A
and right key respectively. known plaintext attack is a cryptanalytic attack in which
The decryption sensitivities with respect to fractional the cryptanalyst possess a substantial quantity of plain
orders are the most significant factors and have been images and corresponding cipher images. A chosen
investigated by calculating MSE between original cipher text attack is a cryptanalytic attack in which the
primary image and decompressed–decrypted image.32,33 cryptanalyst can submit an unlimited number of plain
We observed that sensitivities of encryption keys were images of his own choice and examine the resulting
more for without compression method. If r(i,j) and o(i,j) cipher images.
are the retrieved and original images respectively at (i,j) Several methods using Rivest–Shamir–Adelman
pixel, the MSE can be defined by equation (4) as the (RSA) algorithm and double random phase encoding
energy of the difference image are considered37,38 to manage the security of double
2 random phase encoding. The RSA algorithm in Ref. 37
1 X M X N
improved the signal to noise ratio (SNR) and increase
MSE~kr{ok2 ~ jrði,j Þ{oði,j Þj (4)
M|N i~1 j~1 robustness when compared with.39 The SNR calculated
by Yuan et al. is 5?66 dB. The SNR calculated with RSA
The curves for the MSE versus the changes of algorithm by Saini and Sinha37 is 304?73 dB. Therefore,
fractional orders, –a3 and –a1 have been calculated and it has been demonstrated in Ref. 37 that improvement
depicted in Fig. 11a and b for joint compression– of 299?07 dB is obtained. It has been also demons-
encryption method. In decryption, MSE is minimum trated elsewhere37 that the RSA algorithm37 increased
robustness for wrong decryption key. The SNR of Yuan 12. Li, X. X. and Zhao, D. M. Optical image encryption with
simplified fractional hartley transform, Chin. Phys. Lett., 2008,
et al. with wrong key remains the same and equal to
25, 2477–2480.
5?66 dB. The SNR of Ref.37, obtained with wrong key 13. Wu, J. H., Zhang, L. and Zhou, N. R. Image encryption based on
decryption was 0?17 dB. the multiple-order discrete fractional cosine transform, Opt.
An analysis in Ref. 38 demonstrates that this method Commun., 2010, 283, (9), 1720–1725.
improves security from Ref. 40, for information hiding 14. Zhou, N. R., Dong, T. J. and Wu, J. H. Novel image encryption
algorithm based on multiple-parameter discrete fractional random
using double random phase encoding. In Zhou’s transform, Opt. Commun., 2010, 283, (15), 3037–3042.
method, the technical metrics are quite different from 15. Hennely, B. and Sheridan, J. T. Fractional Fourier transform based
those of composed images. It increases insecurity. image encryption: Phase retrieval algorithm, Opt. Commun., 2003,
Therefore, in Ref. 38 an improved method (with random 226, 61–80.
perturbations) for effective image quality and security is 16. Sinha, A. and Singh, K. Image encryption by using fractional
Fourier transform and jigsaw transform in image bit planes, SPIE-
proposed. The three technical metrics are improved but Int. Soc. Opt. Eng., 2005, 44, 15–18.
worth of composed image has been degraded. An 17. Wang, X., Zhao, D. and Chen, L. Image encryption based on
additional method in Ref. 38, called further improved extended fractional Fourier transform and digital holography
method is proposed and worth of composed image is technique, Opt. Commun., 2006, 260, 449–453.
18. Zhu, B. and Liu, S. Optical image encryption based on multi
kept acceptable.
fractional Fourier transform, Opt. Lett., 2000, 25, 1159–1161.
19. Liu, Z. and Liu, S. Random fractional Fourier transform, Opt.
Conclusion Lett., 2007, 32, 2088–2090.
20. Liu, Z., Zhao, H. and Liu, S. A discrete fractional random
A joint algorithm for image compression and encryption transform, Opt. Commun., 2005, 255, 357–365.
based on DFrFT for compression and DFrCT for 21. Z. Liu and S. Liu: Randomization of the Fourier transform, Opt.
encryption is proposed. It has been observed that Lett., 2007, 32, 478–480.
DFrFT and DFrCT make the full use of the additional 22. Liu, Z., Ahmad, M. A. and Liu, S. A discrete fractional angular
transform, Opt. Commun., 2008, 281, 1424–1429.
degrees of freedom provided by fractional orders to 23. Dang, P. P. and Chau, P. M. Image encryption for secure internet
achieve an optimum domain in compression and multimedia applications, IEEE Trans. Consum. Electron., 2000, 46,
encryption. With change in fractional order value in – (3), 395–403.
a3 and –a1, the value of MSE changes, which shows 24. Ozaktas, H. M., Zalevsky, Z. and Kutay, M. A. The Fractional
Fourier Transform with Applications in Optics and Signal
sensitivity of right keys. Decryption results are more
Processing, 2000 (John Wiley & Sons Ltd, New York).
sensitive for change in –a3 as compared to –a1. The 25. Almeida, B. L. The fractional Fourier transform and time-
proposed joint algorithm has an excellent feature of frequency representations’, Proc. IEEE Trans. Signal Process.,
robustness in security. Numerical simulations verified 1994, 42, 3084–3091.
the validity and efficacy of algorithm. 26. Santhanam, B. and McClellan, J. H. The DRFT-A rotation in time
frequency space, Proc. ICASSP, 1995, 5, 921–924.
27. Pei, S. C. and Ding, J. J. Closed form discrete fractional and affine
Acknowledgement Fourier transforms, IEEE Trans. Signal Process., 2000, 48, 1338–
1353.
The authors would like to thank anonymous reviewers 28. Pei, S. C. and Yeh, M. H. The discrete fractional cosine and sine
for their valuable comments. transforms, IEEE Trans. Signal Process., 2001, 49, 1198–1207.
29. Singh, K. Performance of discrete fractional Fourier transform
classes in signal processing applications, PhD thesis, Thapar
References University, Patiala, 2006.
1. Cebrail, T. and Serder, S. An overview of image compression 30. Saha, S. Image compression-from DCT to Wavelets: a review, The
approaches, Proc. 3rd Int. Conf. on Digital Telecommunication, ACM Student Magazine, 2000.
Koblenz, Germany, December 2008, pp. 164–208. 31. Tao, R., Xin, Y. and Wang, Y. Double image encryption based on
2. Yetik, I. S., Kutay, M. A. and Ozaktas, H. M. Image representa- random phase encoding in the fractional Fourier domain, Opt.
tion and compression with the fractional Fourier transform, Opt. Express, 2007, 15, 16067–16079.
Commun., 2001, 197, 275–278. 32. Liu, S., Yu, L. and Zhu, B. Optical image encryption by cascaded
3. Alfalou, A., Elbouz, M., Mansour, A. and Keryer, G. New spectral fractional Fourier transform with random phase filtering, Opt.
image compression method based on an optimal phase coding and Commun., 2001, 187, 57–63.
the RMS duration principle, J. Opt., 2010, 12, 115403. 33. Ozturk, I. and Sogukpinar, I. Analysis and comparison of Image
4. Liu, Z., Li, Q., Dai, J., Sun, X., Liu, S. and Ahmad, M. A. A new Encryption Algorithms, World Acad. Sci., Eng. Technol., 2005, 3,
kind of double image encryption by using a cutting spectrum in the 26–30.
1-D fractional Fourier transform domains, Opt. Commun., 2009, 34. Jin, W. and Yan, C. Optical image encryption based on multi-
282, 1536–1540. channel fractional Fourier transform and double random phase
5. Hennely, B. and Sheridan, J. T. Image encryption and fractional encoding technique, Optik, 2007, 118, 38–41.
Fourier transform, Optik, 2003, 114, 251–265. 35. Sheng, Y., Xin, Z., Alam, M. S., Xi, L. and Xiao-Feng, L.
6. Gonzalez, R. C. and Woods, R. E. Digital Image Processing, 2004, Information hiding based on double random-phase encoding and
2nd edition (Pearson Education (Singapore) Pte. Ltd, Delhi). public-key cryptography, Opt. Express, 2009, 17, (5), 3270–3284.
7. Vijaya, C. and Bhat, J. S. Signal compression using discrete 36. Frauel, Y., Castro, A., Naughton, T. J. and Javidi, B. Resistance of
fractional Fourier transform and set partitioning in hierarchical the double random phase encryption against various attacks, Opt.
tree, Signal Process., 2006, 86, 1976–1983. Express, 2007, 15, (6), 10253–10265.
8. Zhou, N. R., Wang, Y. X., Gong, L. H., He, H. and Wu, J. H. 37. Saini, N. and Sinha, A. Key management of the double random-
Novel single-channel color image encryption algorithm based on phase-encoding method using public-key encryption, Opt. Lasers
chaos and fractional Fourier transform, Opt. Commun., 2011, 284, Eng., 2010, 48, 329–334.
(12), 2789–2796. 38. Sang, J., Xiang, H., Fu, L. and Sang, N. Security analysis and
9. Singh, N. and Sinha, A. Optical image encryption using improper improvement on a double-random phase-encoding technique based
Hartley transforms and chaos, Optik, 2010, 121, (10), 918–925. information hiding method, Opt. Commun., 2009, 282, 2307–2317.
10. Zhou, N. R., Wang, Y. X. and Gong, L. H. Novel optical image 39. Yuan, S., Zhou, X., Li, D.-H. and Zhou, D.-F. Simultaneous
encryption scheme based on fractional Mellin transform, Opt. transmission for an encrypted image and a double random-phase
Commun., 2011, 284, (13), 3234–3242. encryption key, Appl. Opt., 2007, 46, 3747–3753.
11. Zhou, N. R., Wang, Y. X. and Wu, J. H. Image encryption 40. Zhou, X., Lai, D., Yuan, S., Li, D.-H. and Hu, J.-P. A method for
algorithm based on the multi-order discrete fractional Mellin hiding information utilizing double-random phase-encoding tech-
transform, Opt. Commun., 2011, 284, (24), 5588–5597. nique, Opt. Laser Technol., 2007, 39, (7), 1360–1363.