Learning-based attacks for detecting the vulnerability of computer-generated hologram based optical encryption

Zhou, Lina; Xiao, Yin; Chen, Wen

doi:10.1364/oe.380004

Cited by 37 publications

(13 citation statements)

References 45 publications

(64 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Reconstructing images through scattering media with highly scalable CNN approaches , have been demonstrated to exhibit optical encryption transmission. Different from some studies , on using the CNN in the decryption process, which might suffer from the known-plaintext attacks (KPA), our designed transfer system has very high security due to the fact that the diffuser is dynamically shaken instead of being steady. Traditional phase retrieval or transfer matrix calibration techniques can hardly recover the encoded information when the diffuser is changing randomly with time.…”

Section: Discussionmentioning

confidence: 99%

Deep Learning-Enabled Orbital Angular Momentum-Based Information Encryption Transmission

Feng

Guo

et al. 2022

ACS Photonics

View full text Add to dashboard Cite

Orbital angular momentum (OAM)-based optical encryption transmission plays an important role in optical communications. However, it remains challenging to encrypt the data with great security and decrypt them with high fidelity while maintaining large-capacity transmission. In this paper, we propose a novel optical encryption transmission technique, which dynamically converts OAM modes into speckle patterns with a randomly shaking diffuser for high-security encryption, and a pretrained convolutional neural network (CNN) is later employed to extract encoded information hidden in the speckle patterns for high-fidelity information decryption. Our experiment demonstrates that the modulated OAM with an interval of a topological charge as small as 0.01 can be recognized by the CNN with an accuracy of 99.83%. To demonstrate its application, a cat image, encoded and encrypted by the designed system, has been successfully decrypted and decoded with a bit error rate of 0.008%. The security mechanism of the technique has also been experimentally verified and discussed. This technique thereby provides a new avenue for OAM-based encrypted optical information transmission.

show abstract

Section: Discussionmentioning

confidence: 99%

Deep Learning-Enabled Orbital Angular Momentum-Based Information Encryption Transmission

Feng

Guo

et al. 2022

ACS Photonics

View full text Add to dashboard Cite

show abstract

“…The security of the CDPE scheme itself has been sufficiently investigated in [9][10][11], so we will focus on the security enhancement contributed by the compression procedure. According to the Kerckhoffs' assumption [30], the attacker has the total knowledge of the cryptosystem except the secret keys.…”

Section: Security Analysismentioning

confidence: 99%

“…Among the existing optical encryption schemes, the ones that encrypt the primary image into phase-only masks (POMs) are quite attractive, as the decoded image can be directly captured by the intensity sensitive devices in such schemes [5,8,9]. In particular, a cascaded double phase encryption (CDPE) scheme is worthy of noting due to its high security [9], and some variants of it were also reported and studied recently [10,11]. In addition to developing new optical cryptosystems, people also make efforts to compress the ciphertext to facilitate its transmission and storage [12].…”

Section: Introductionmentioning

confidence: 99%

Optical Compressive Encryption via Deep Learning

Qin

Wan

et al. 2021

IEEE Photonics J.

View full text Add to dashboard Cite

The compression of the ciphertext of a cryptosystem is desirable considering the dramatic increase in secure data transfer via Internet. In this paper, we propose a simple and universal scheme to compress and decompress the ciphertext of an optical cryptosystem by the aid of deep learning (DL). For compression, the ciphertext is first resized to a relatively small dimension by bilinear interpolation and thereafter condensed by the JPEG2000 standard. For decompression, a well-trained deep neural network (DNN) can be employed to perfectly recover the original ciphertext, in spite of the severe information loss suffered by the compressed file. In contrast with JPEG2000 and JPEG, our proposal can achieve a far smaller size of the compressed file (SCF) while offering comparable decompression quality. In addition, the SCF can be further reduced by compromising the quality of the recovered plaintext. It is also shown that the compression procedure can provide an additional security level, and this may offer new insight into the compressive encryption in optical cryptosystems. Both simulation and experimental results are presented to demonstrate the proposal.

show abstract

“…Recently, deep learning (DL) has achieved widespread use in various fields, such as image denoising, image inpainting, natural language processing, to name a few 15 . DL has also made a comprehensive application in computational imaging 16 such as digital holography 17 , lensless imaging 18 , imaging through scattering media and turbid media 19 – 21 , optical encryption 22 , 23 and ghost imaging 24 – 28 . The term of ghost imaging using deep learning (GIDL) was first proposed by Lyu et al 24 .…”

Section: Introductionmentioning

confidence: 99%

A residual-based deep learning approach for ghost imaging

Bian

et al. 2020

Sci Rep

View full text Add to dashboard Cite

Ghost imaging using deep learning (GiDL) is a kind of computational quantum imaging method devised to improve the imaging efficiency. However, among most proposals of GIDL so far, the same set of random patterns were used in both the training and test set, leading to a decrease of the generalization ability of networks. Thus, the GIDL technique can only reconstruct the profile of the image of the object, losing most of the details. Here we optimize the simulation algorithm of ghost imaging (GI) by introducing the concept of "batch" into the pre-processing stage. It can significantly reduce the data acquisition time and create reliable simulation data. The generalization ability of GIDL has been appreciably enhanced. Furthermore, we develop a residual-based framework for the GI system, namely the double residual U-Net (DRU-Net). The imaging quality of GI has been tripled in the evaluation of the structural similarity index by our proposed DRU-Net. Ghost imaging (GI) is an unconventional imaging method compared with traditional optical imaging methods. It was first proposed as a quantum entangled phenomenon by making use of the entangled two-photon light source generated by the spontaneous parameter down-conversion (SPDC) process 1. However, GI was demonstrated to be accomplished by classical incoherent light source soon later 2. It led to controversy focus on whether the quantum entangled is necessary for the GI system. The GI system contains two distributed light beams, which are the test beam and the reference beam. In the test beam, the light illuminates the object directly and then be collected into a bucket measurement. In the reference beam, the light travels freely to a high-resolution detector without interacting with the object. The image can then be reconstructed through the correlation measurement between the two light beam signals 3. In terms of computational ghost imaging (CGI), the reference beam becomes unnecessary as one can obtain the image by calculating the correlation between the test beam intensity and the knowledge of the random patterns displayed on the spatial light modulator (SLM) 4-6. CGI usually needs a large number of random patterns to avoid noise disturbance to achieve images of high resolution. This requirement leads to a long data acquisition time, which is the main issue preventing CGI from practical use. Improved correlation method such as normalized ghost imaging 7 and differential ghost imaging 8 was later proposed to increase imaging efficiency. Notably, the Gerchberg-Saxton algorithm and compressive sensing ghost imaging (CSGI) regard GI as an optimization problem. In 9 , the Gerchberg-Saxton-like technique takes the integral property of Fourier transform into full consideration to provide a different perspective of image reconstruction of GI. CSGI can reduce measurement times and boost imaging quality 10-14. Recently, deep learning (DL) has achieved widespread use in various fields, such as image denoising, image inpainting, natural language processing, to name a few 15. DL h...

show abstract

Learning-based attacks for detecting the vulnerability of computer-generated hologram based optical encryption

Cited by 37 publications

References 45 publications

Deep Learning-Enabled Orbital Angular Momentum-Based Information Encryption Transmission

Deep Learning-Enabled Orbital Angular Momentum-Based Information Encryption Transmission

Optical Compressive Encryption via Deep Learning

A residual-based deep learning approach for ghost imaging

Contact Info

Product

Resources

About