A new kind of double image encryption by using a cutting spectrum in the 1-D fractional Fourier transform domains
Introduction
In the recent years, many optical image encryption algorithms [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20] based on some tools or models, such as double random phase encoding, digital holography, phase retrieval, intensity random coding and virtual optics, have been proposed. A classical one of them is random phase encoding [1], which has been proved that the security is attractive in the aspect against some attacks [21], [22]. Moreover, this algorithm has been developed into the domains of fractional Fourier transform (FRFT) [23] and Fresnel transform [24]. Recently, the FRFT and its improved versions have been used in the field of image encryption [16], [17], [18], [19], [20]. In fact, the former algorithms are researched for encrypting single image with some random elements. However, the problem that how to hide more than one image was ignored in a long time.
Multiple image encryption (MIE) is a new concept and interesting in the field of optical information security. There multiple 2-D data, such as a series of pictures of an object or many different images, is encrypted with certain relations, which can be often stored or transmitted together. Situ and Zhang first proposed an algorithm of MIE by using wavelength multiplexing [25]. Because the multiple original images are directly added into one image, the qualities of the corresponding decrypted images in the algorithms [25] are not perfect due to the cross-talk effects between images. Four kinds of double image encryption schemes [18], [26], [27], [28] were reported, which belong to the algorithm of MIE. One of them can be expanded to encrypt more than two images [26]. However, the phase retrieval algorithm in the paper [26] requires complex computational load.
In this paper, we propose a new kind of double image encryption by use of cutting spectrum in the FRFT domains. The center part of spectrum includes main information of an image, which is considered for describing and recording the image approximatively. The higher frequency of FRFT is cut down and replaced by zero in the algorithm. Although this operation sacrifices the detail information of image, the visual difference between original image and recovered image is very small. The part spectra of two images are scrabbled up a new spectrum, which is encrypted with double random phase encoding technique. All transforms are limited in the format of one-dimension, which was used for optical image encryption [29].
The rest of the paper is organized in the following sequence. In Section 2, the algorithm of double image encryption is described and analyzed in detail. In Section 3, numerical result is given to demonstrate the validity of the algorithm. Some concluding remarks are summarized in the final section.
Section snippets
Image encryption algorithm
The double image encryption algorithm is given by using FRFT and cutting operation. There is 1-D FRFT for image encryption, which is different from some algorithms. The 1-D transforms at two orthogonal directions are employed in the encryption scheme.
Numerical simulation
A series of numerical simulation experiments have been achieved to verify the validity and efficiency of the proposed method and investigate its performance. The fractional order is for cutting operation and image encryption. The fractional order is fixed 0.8. Two images with pixels are used for the simulation and represented in Fig. 3a and b, which are grey-level image and binary image, respectively. The CC values of Fig. 3a and b under the corresponding parameters used in
Conclusion
We have proposed a new algorithm of double image encryption based on 1-D FRFT. In this algorithm, two center parts of FRFT spectrum of original images are combined together into a spectrum image by use of 1-D FRFT at x-direction. Double random phase encoding technology and 1-D FRFT are utilized for image encryption. Although the detail information of image is deleted by cutting operation, the evident advantages of the proposed scheme are small computational load and the elimination of
Acknowledgements
This work was supported by the National Natural Science Foundation of China under Grant Nos. 10674038 and 10604042 and National Basic Research Program of China under Grant 2006CB302901, China Postdoctoral Science Foundation (20080430913), and development program for outstanding young teachers in Harbin Institute of Technology, HITQNJS. 2008. 027. The authors greatly appreciate the constructive comments and suggestions by the anonymous reviewers.
References (29)
- et al.
Opt. Commun.
(2005) - et al.
Opt. Commun.
(2007) - et al.
Opt. Commun.
(2006) - et al.
Opt. Commun.
(2008) - et al.
Opt. Commun.
(2007) - et al.
Opt. Commun.
(2002) - et al.
Optik
(2003) - et al.
Opt. Commun.
(2009) - et al.
Opt. Commun.
(2007) - et al.
Opt. Commun.
(2007)
Opt. Lett.
Opt. Lett.
Appl. Opt.
Opt. Exp.
Cited by (53)
Vulnerability to machine learning attacks of optical encryption based on diffractive imaging
2020, Optics and Lasers in EngineeringStego optical encryption based on chaotic S-box transformation
2014, Optics and Laser TechnologyDouble image encryption scheme by using random phase encoding and pixel exchanging in the gyrator transform domains
2013, Optics and Laser TechnologyFour-image encryption method based on spectrum truncation, chaos and the MODFrFT
2013, Optics and Laser TechnologyCitation Excerpt :This method can encrypt more images simultaneously by cutting less central spectrum, but the qualities of decrypted images will be worse, for the higher frequency coefficients are lost in the process of spectrum truncating. Other multiple-image encryption methods based on spectrum truncation have similar drawbacks [20–23]. In many cases, a series of pictures of an object or many different images are encrypted with certain relations, which can be often stored or transmitted together.