Short communicationLight field image encryption based on spatial-angular characteristic
Introduction
With the rapid development of computer network and multimedia technology, more and more information is transmitted through the network, and information security is facing increasingly severe challenges. As a typical multimedia resource, the digital image is faced with increasingly prominent security problems. Light field (LF) imaging, as emerging imaging technology, can simultaneously capture the intensity and direction of light propagation, and has attracted more and more attention. Meanwhile, LF have been widely developed in depth estimation [1], compression [2], and quality assessment [3] in recent years. The increasing commercialization of LF cameras such as Picam [4] and Lytro [5] has led to their further popularity. However, there are still few works on the security encryption of LF images. The security protection of LF images need further attention.
Encryption is the most direct and basic method to protect digital image information security. At present, image encryption technologies can be roughly divided into chaotic system [6], DNA sequences [7], compressive sensing [8], and optical image encryption [9]. These types of methods have their own advantages and disadvantages in terms of encryption performance. In [10], Ismail et al. presented a novel lossless image encryption algorithm based on edge detection and generalized chaotic maps for key generation. Wang et al. designed a visually secure encryption scheme by using the parallel compressive sensing (PCS) counter mode and embedding technique [11]. Optical encryption techniques can use the amplitude, phase, and wavelength of light to encode images, which have high speed in parallel processing of 2-D information.
Refregier and Javidi firstly adopted two random phase plates to design double random phase encoding (DRPE) system [12]. The system takes two identical Fourier lenses as the main body, and their distance is the focal length f. The DRPE system has the distinct advantage of parallel processing of two-dimensional complex data and high speed, but its disadvantage is that the degree of freedom of key space is small. Based on the Fresnel diffraction theory, Situ et al. proposed an image encryption method that can achieve efficient encryption [13], but the method occupied substantial computing resources and storage resources. Follow Refregier’s work, Unnikrishnan et al. introduced the fractional Fourier transform (FrFT) theory to design an image encryption scheme, where the lens parameters are used as part of the key to expand the key space [14]. In [15], the authors applied digital holography technology to double random phase encoding, and employ the Joint Transform Correlator (JTC) to build a security system that can be used for video storage and transmission. Zhong et al. proposed a novel approach using discrete multiple-parameter FrFT for image encryption based on interference [16].
Image encryption methods based on optics have been continuously proposed, and various theories have been introduced into the field of image encryption. However, these methods mainly considered the encryption for traditional images. At the same time, there are still little research on LF image encryption scheme. You et al. proposed a micro-lens to record the spatial and directional information for 3-D image encryption [17]. Wen et al. focused on color LF image encryption based on DNA sequence and chaotic system [18]. The authors reported a novel optical encryption strategy that utilizes highly scattered wavefront of LF to encrypt the plaintext and exploits a scattering medium as the unique physical key in Liu et al. [19]. In summary, these methods only explore the spatial information of the LF image and do not fully consider the angular information of LF image.
Different from the existing LF image encryption methods, this paper simultaneously considers the internal spatial-angular characteristic of the LF image, and propose a novel encryption scheme for LF image. Specifically, the discrete fractional Fourier transform (DFrFT) is adopted to extract specific spatial features of LF image. Then, the epipolar plane image (EPI) is generated from multiple views of the LF image. The ciphertext EPI obtained by Arnold transform is displayed as an angular domain ciphertext image after reverse reconstruction. Finally, weighted linear blending and Chen’s chaotic system are performed on the ciphertext images in the spatial and angular domains to obtain the final ciphertext image. The experimental results demonstrate the effectiveness and scalability of the proposed scheme for LF image encryption.
The rest of this paper is organized as follows. Section 2 introduces the related work. Section 3 describes the proposed LF image encryption scheme. Experimental results and analysis are presented in Section 4. Section 5 concludes this paper and discusses future work.
Section snippets
Related work
As a novel imaging technology, LF has attracted extensive attention because it can capture richer visual information in space [20]. In contrast with conventional camera imaging devices, an array of micro-lens arrays is incorporated in the LF imaging device, which can obtain rich visual perception information in 3-D space. In Fig. 1, the LF imaging device is mainly composed of the main lens, a micro-lens array, and an image sensor. The main lens is used to collect light, and the micro-lens array
The proposed scheme
The LF model is composed of two two-dimensional planes, namely the angular domain and the spatial domain. The spatial domain is represented by the LF different sub-view images, which contains richer content information, while the angular domain reflects the variation of the different views. The representativeness of the features represented by the spatial and angular domains within the LF are different and their importance varies. Therefore, it is very meaningful to construct a suitable
Experimental results
In order to measure the security performance of the proposed encryption method, we conducted sufficient experiments in two categories: statistical analysis and security analysis. Statistical analysis contains histogram analysis, information entropy analysis and correlation analysis. In addition, security analysis covers key space analysis, resistance to differential attack analysis, key sensitivity analysis and encryption efficiency analysis [25], [26]. Table 1 shows the comparison of our
Conclusion and future work
In summary, an LF image encryption strategy based on spatial-angular characteristic is proposed. The LF image contains both spatial and angular domain features. This suggests that, the sub-view image is represented as the spatial domain of the LF, which is encrypted by using DFrFT. In addition, an EPI is constructed by extracting pixels from multiple sub-view images as angular domain information, and the cipher EPI is obtained by employing Arnold transform as the angular domain ciphertext
CRediT authorship contribution statement
Kangkang Wei: Conceptualization, Methodology, Software, Writing - original draft. Wenying Wen: Data curation, Investigation, Writing - original draft. Yuming Fang: Visualization, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by the National Key R&D Program of China (Grant 2018AAA0100601), the Natural Science Foundation of China (Grant nos. 61961022, 61822109, 61961021), the Training Program for Academic and Technical Leaders of Jiangxi Province (Grant no. 20204BCJL23036), the Natural Science Foundation of Jiangxi Province (Grant no. 20202BAB202022), and Postdoctoral Research Project of Jiangxi Province (Grant no. 2020KY14 ).
References (32)
- et al.
A color image cryptosystem based on dynamic DNA encryption and chaos
Signal Process.
(2019) - et al.
A visually secure image encryption scheme based on semi-tensor product compressed sensing
Signal Process.
(2020) - et al.
Edge-based lightweight image encryption using chaos-based reversible hidden transform and multiple-order discrete fractional cosine transform
Opt. Laser Technol.
(2013) - et al.
A novel image encryption system merging fractional-order edge detection and generalized chaotic maps
Signal Process.
(2020) - et al.
A visually secure image encryption scheme based on parallel compressive sensing
Signal Process.
(2019) - et al.
Micro-lens array based 3-D color image encryption using the combination of gravity model and Arnold transform
Opt. Commun.
(2015) - et al.
Optical color image encryption based on Hadamard single-pixel imaging and Arnold transformation
Opt. Lasers Eng.
(2021) - et al.
An image encryption scheme based on hybridizing digital chaos and finite state machine
Signal Process.
(2019) - et al.
Cross-plane colour image encryption using a two-dimensional logistic tent modular map
Inf. Sci.
(2021) - et al.
Perceptual image hashing via dual-cross pattern encoding and salient structure detection
Inf. Sci.
(2018)
Unsupervised monocular depth estimation from light field image
IEEE Trans. Image Process.
Content-based light field image compression method with gaussian process regression
IEEE Trans. Multimed.
Light field image quality assessment by local and global features of epipolar plane image
IEEE International Conference on Multimedia Big Data
Picam: an ultra-thin high performance monolithic camera array
ACM Trans. Graph. Siggraph Asia
Color image encryption based on joint permutation and diffusion
J. Electron. Imaging
Cited by (8)
Light field image encryption based on steerable cosine number transform
2023, Signal ProcessingCitation Excerpt :In this context, none of the published works, with respect to encryption, exploit the whole 4D structure. Most of them exploit only the spatial plane [11–15] and only one of them, to the best of our knowledge, also explores the angular plane [9], however, performing the encryption in each of the views separately. As far as we are concerned, six papers were published on this context.
Securing color image by using bit-level modified integer nonlinear coupled chaos model in Fresnel diffraction domains
2022, Optics and Lasers in EngineeringAn effective decomposition-enhancement method to restore light field images captured in the dark
2021, Signal ProcessingCitation Excerpt :Compared with what can be done by conventional cameras, light field (LF) cameras with micro-lens array can capture both the intensities and directions of the light rays [1–4] and produce 4D LF images consisting of multiple sub-aperture images (SAIs) to record the scene from different directions.
A new encryption model for multimedia content using two dimensional Brownian motion and coupled map lattice
2023, Multimedia Tools and ApplicationsAn image encryption scheme based on multi-objective optimization and block compressed sensing
2022, Nonlinear Dynamics