Chapter

Image Encryption and Compression Based on a VAE Generative Model: 4th International Conference, ICCCS 2018, Haikou, China, June 8–10, 2018, Revised Selected Papers, Part III

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

To solve the problem that the network security real-time transmits image, a new image encryption and compression method based on a variational auto-encoder (VAE) generative model is proposed in this paper. The algorithm aims to encrypt and compress images by using a variational auto-encoder generative model. Firstly, we use multi-layer perceptual neural network to train the VAE model, and set parameters of the model to get the best model. Then, the peak signal-to-noise ratio (PSNR) and mean square error (MSE) are used to measure the compression effect and Set the number of iterations of the model. Finally, we extract the data of based on a variational auto-encoder and perform division, then the data input the VAE generative model to encrypt image and analyze encryption images. In this paper, we use the standard image of 256 * 256 to do simulation experiments and use histogram and image correlation to analyze the results of encryption. The simulation results show that the proposed method can effectively compress and encrypt images, and then obtain better compression image than stacked auto-encoder (SAE), while the algorithm is faster and easier encrypting and decrypting images and the decrypted image distortion rate is low and suitable for practical applications.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
We describe an image compression method, consisting of a nonlinear analysis transformation, a uniform quantizer, and a nonlinear synthesis transformation. The transforms are constructed in three successive stages of convolutional linear filters and nonlinear activation functions. Unlike most convolutional neural networks, the joint nonlinearity is chosen to implement a form of local gain control, inspired by those used to model biological neurons. Using a variant of stochastic gradient descent, we jointly optimize the entire model for rate-distortion performance over a database of training images, introducing a continuous proxy for the discontinuous loss function arising from the quantizer. Under certain conditions, the relaxed loss function may be interpreted as the log likelihood of a generative model, as implemented by a variational autoencoder. Unlike these models, however, the compression model must operate at any given point along the rate-distortion curve, as specified by a trade-off parameter. Across an independent set of test images, we find that the optimized method generally exhibits better rate-distortion performance than the standard JPEG and JPEG 2000 compression methods. More importantly, we observe a dramatic improvement in visual quality for all images at all bit rates, which is supported by objective quality estimates using MS-SSIM.
Article
Full-text available
With the evolution of technologies, the size of an image data has been significantly increased. However, traditional image encryption schemes cannot handle the emerging problems in big data such as noise toleration and compression. In order to meet today’s challenges, we propose a new image encryption scheme based on chaotic maps and orthogonal matrices. The main core of the proposed scheme is based on the interesting properties of an orthogonal matrix. To obtain a random orthogonal matrix via the Gram Schmidt algorithm, a well-known nonlinear chaotic map is used in the proposed scheme to diffuse pixels values of a plaintext image. In the process of block-wise random permutation, the logistic map is employed followed by the diffusion process. The experimental results and security analyses such as key space, differential and statistical attacks show that the proposed scheme is secure enough and robust against channel noise and JPEG compression. In addition to complete encryption for higher security, it also supports partial encryption for faster processing as well.
Article
Full-text available
In this paper, the quaternion discrete Cosine transform (QDCT) is applied to the color image encryption technology, and partial encryption technique of color image based on QDCT is proposed. Obtained the QDCT spectrum of color image by QDCT, the real and imaginary parts of QDCT coefficients are quantized and encoded with the color image quantization coefficient table and QDCT coefficients encoding method, and part of low and high frequency coefficients are encrypted in the encoding process. Simulation results and security performance analysis shows that the proposed scheme can obtain the effect of secure image encryption, reduce the amount of encryption data and transmission data, and well distribute the pixels of the original image when only part of the data are encrypted. What's more, the proposed scheme can decrypt and recover the original image with high quality. Due to the large key space, the proposed scheme can resist exhaustive attacks.
Article
Full-text available
In this paper we propose noise removal approach using spatial gradient based bilateral filter and minimum mean square error filtering. The proposed method consist two-steps process. In first step, we generate a reference image from the given noisy image, by extracting (2W + 1) ×(2W + 1)size patches from it and then apply proposed spatial gradient based bilateral filter on each of these patches. To reduce the mean square error, in second step we apply minimum mean square error (MSE) filter on the reference image. Generally all the noise removal approaches change the natural appearance of the restored image, while the proposed method restores the image without affecting its natural appearance.
Article
Full-text available
Recent advances in stochastic variational inference have made it possible to construct variational posterior approximations containing auxiliary random variables. This enables us to explore a new synthesis of variational inference and Monte Carlo methods where we incorporate one or more steps of MCMC into our variational approximation. This note describes the theoretical foundations that make this possible and shows some promising first results.
Article
Full-text available
We propose develop a generalized algorithm for hiding audio signal using image steganography. The authors suggest transmitting short audio messages camouflaged in digital images using Principal Component Analysis as an encryption technique. The quantum of Principal components required to represent the audio signal by removing the redundancies is a measure of the magnitude of the Eigen values. The aforementioned technique follows a dual task of encryption and in turn also compresses the audio data, sufficient enough to be buried in the image. A 57Kb audio signal is decipher from the Stego image with a high PSNR of 47.49 and a correspondingly low MSE of 3.3266×10 -6 with an equalized high quality audio output.
Conference Paper
Full-text available
In this paper, we analyse two well-known objective image quality metrics, the peak-signal-to-noise ratio (PSNR) as well as the structural similarity index measure (SSIM), and we derive a simple mathematical relationship between them which works for various kinds of image degradations such as Gaussian blur, additive Gaussian white noise, jpeg and jpeg2000 compression. A series of tests realized on images extracted from the Kodak database gives a better understanding of the similarity and difference between the SSIM and the PSNR.
Article
The mainstream adaptive steganography algorithms often cannot transmit secret messages correctly when stego images suffer from JPEG compression. In this respect, researchers proposed a series of robust adaptive steganography methods based on the framework of “Compression-resistant Domain Constructing + RS-STC Codes” in previous studies. However, these methods leave behind the fault tolerance analysis, resulting in potential mistakes in extracted messages, which brings uncertainty to practical application. To solve this problem, an error model based on burst errors and STCs decoding damage is given in this manuscript, utilizing the burst error model based on Poisson distribution. Then the model is verified using the hypothesis test problem judged by the χ² test method. Based on the proposed model, the error conditions of received stego sequence are depicted, and the fault-tolerant performance of the robust steganography based on “Compression-resistant Domain Constructing + RS-STC Codes” is deduced, that is, the probability lower bound for RS-STCs decoding to correctly extract embedded messages. Experiments demonstrate that the practical fault-tolerant results of previous robust steganography methods consist with the theoretical derivation results, which provides a theory support for coding parameter selection and message extraction integrity to the robust steganography based on “Compression-resistant Domain Constructing + RS-STC Codes”.
Conference Paper
This paper introduces the Deep Recurrent Attentive Writer (DRAW) neural network architecture for image generation. DRAW networks combine a novel spatial attention mechanism that mimics the foveation of the human eye, with a sequential variational auto-encoding framework that allows for the iterative construction of complex images. The system substantially improves on the state of the art for generative models on MNIST, and, when trained on the Street View House Numbers dataset, it generates images that cannot be distinguished from real data with the naked eye.
Conference Paper
We marry ideas from deep neural networks and approximate Bayesian inference to derive a generalised class of deep, directed generative models, endowed with a new algorithm for scalable inference and learning. Our algorithm introduces a recognition model to represent approximate posterior distributions, and that acts as a stochastic encoder of the data. We develop stochastic back-propagation -- rules for back-propagation through stochastic variables -- and use this to develop an algorithm that allows for joint optimisation of the parameters of both the generative and recognition model. We demonstrate on several real-world data sets that the model generates realistic samples, provides accurate imputations of missing data and is a useful tool for high-dimensional data visualisation.
Article
We describe an image compression system, consisting of a nonlinear encoding transformation, a uniform quantizer, and a nonlinear decoding transformation. Like many deep neural network architectures, the transforms consist of layers of convolutional linear filters and nonlinear activation functions, but we use a joint nonlinearity that implements a form of local gain control, inspired by those used to model biological neurons. Using a variant of stochastic gradient descent, we jointly optimize the system for rate-distortion performance over a database of training images, introducing a continuous proxy for the discontinuous loss function arising from the quantizer. The relaxed optimization problem resembles that of variational autoencoders, except that it must operate at any point along the rate-distortion curve, whereas the optimization of generative models aims only to minimize entropy of the data under the model. Across an independent database of test images, we find that the optimized coder exhibits significantly better rate-distortion performance than the standard JPEG and JPEG 2000 compression systems, as well as a dramatic improvement in visual quality of compressed images.
Article
We propose a new approach to the problem of optimizing autoencoders for lossy image compression. New media formats, changing hardware technology, as well as diverse requirements and content types create a need for compression algorithms which are more flexible than existing codecs. Autoencoders have the potential to address this need, but are difficult to optimize directly due to the inherent non-differentiabilty of the compression loss. We here show that minimal changes to the loss are sufficient to train deep autoencoders competitive with JPEG 2000 and outperforming recently proposed approaches based on RNNs. Our network is furthermore computationally efficient thanks to a sub-pixel architecture, which makes it suitable for high-resolution images. This is in contrast to previous work on autoencoders for compression using coarser approximations, shallower architectures, computationally expensive methods, or focusing on small images.
Article
Stacked Auto-Encoder (SAE) is a kind of deep learning algorithm for unsupervised learning. Which has multi layers that project the vector representation of input data into a lower vector space. These projection vectors are dense representations of the input data. As a result, SAE can be used for image compression. Using chaotic logistic map, the compression ones can further be encrypted. In this study, an application of image compression and encryption is suggested using SAE and chaotic logistic map. Experiments show that this application is feasible and effective. It can be used for image transmission and image protection on internet simultaneously.
Article
We introduce a simple recurrent variational auto-encoder architecture that significantly improves image modeling. The system represents the state-of-the-art in latent variable models for both the ImageNet and Omniglot datasets. We show that it naturally separates global conceptual information from lower level details, thus addressing one of the fundamentally desired properties of unsupervised learning. Furthermore, the possibility of restricting ourselves to storing only global information about an image allows us to achieve high quality 'conceptual compression'.
Article
Identifying subjects with variations caused by poses is one of the most challenging tasks in face recognition, since the difference in appearances caused by poses may be even larger than the difference due to identity. Inspired by the observation that pose variations change non-linearly but smoothly, we propose to learn pose-robust features by modeling the complex non-linear transform from the non-frontal face images to frontal ones through a deep network in a progressive way, termed as stacked progressive auto-encoders (SPAE). Specifically, each shallow progressive auto-encoder of the stacked network is designed to map the face images at large poses to a virtual view at smaller ones, and meanwhile keep those images already at smaller poses unchanged. Then, stacking multiple these shallow auto-encoders can convert non-frontal face images to frontal ones progressively, which means the pose variations are narrowed down to zero step by step. As a result, the outputs of the topmost hidden layers of the stacked network contain very small pose variations, which can be used as the pose-robust features for face recognition. An additional attractiveness of the proposed method is that no pose estimation is needed for the test images. The proposed method is evaluated on two datasets with pose variations, i.e., MultiPIE and FERET datasets, and the experimental results demonstrate the superiority of our method to the existing works, especially to those 2D ones.
Chapter
Vector quantization (VQ) is a generalization of scalar quantization to the quantization of a vector, an ordered set of real numbers. The jump from one dimension to multiple dimensions is a major step and allows a wealth of new ideas, concepts, techniques, and applications to arise that often have no counterpart in the simple case of scalar quantization. While scalar quantization is used primarily for analog-to-digital conversion, VQ is used with sophisticated digital signal processing, where in most cases the input signal already has some form of digital representation and the desired output is a compressed version of the original signal. VQ is usually, but not exclusively, used for the purpose of data compression. Nevertheless, there are interesting parallels with scalar quantization and many of the structural models and analytical and design techniques used in VQ are natural generalizations of the scalar case.
Article
In this paper, we propose a novel reversible data hiding scheme in encrypted image. The content owner encrypts the original image with the encryption key to achieve privacy protection for image content, and then, each block of the encrypted image is embedded with one secret bit by the data hider using the data-hiding key. Through the elaborate selection for partial pixels to be flipped, data hiding process only conducts slighter modifications to each block, which leads to significant improvement of visual quality for the decrypted image. The receiver can easily decrypt the marked, encrypted image using the encryption key, and then, through the data-hiding key and an adaptive evaluation function of smoothness characteristic along the isophote direction, secret data can be extracted from the decrypted image, and the original image can further be recovered successfully. Experimental results demonstrate the effectiveness of the proposed scheme.
Article
This paper presents the Deep Convolution Inverse Graphics Network (DC-IGN), a model that learns an interpretable representation of images. This representation is disentangled with respect to transformations such as out-of-plane rotations and lighting variations. The DC-IGN model is composed of multiple layers of convolution and de-convolution operators and is trained using the Stochastic Gradient Variational Bayes (SGVB) algorithm. We propose a training procedure to encourage neurons in the graphics code layer to represent a specific transformation (e.g. pose or light). Given a single input image, our model can generate new images of the same object with variations in pose and lighting. We present qualitative and quantitative results of the model's efficacy at learning a 3D rendering engine.
Article
This paper introduces the Deep Recurrent Attentive Writer (DRAW) neural network architecture for image generation. DRAW networks combine a novel spatial attention mechanism that mimics the foveation of the human eye, with a sequential variational auto-encoding framework that allows for the iterative construction of complex images. The system substantially improves on the state of the art for generative models on MNIST, and, when trained on the Street View House Numbers dataset, it generates images that cannot be distinguished from real data with the naked eye.
The ever-increasing size of modern data sets combined with the difficulty of obtaining label information has made semi-supervised learning one of the problems of significant practical importance in modern data analysis. We revisit the approach to semi-supervised learning with generative models and develop new models that allow for effective generalisation from small labelled data sets to large unlabelled ones. Generative approaches have thus far been either inflexible, inefficient or non-scalable. We show that deep generative models and approximate Bayesian inference exploiting recent advances in variational methods can be used to provide significant improvements, making generative approaches highly competitive for semi-supervised learning.
Article
High-dimensional data can be converted to low-dimensional codes by training a multilayer neural network with a small central layer to reconstruct high-dimensional input vectors. Gradient descent can be used for fine-tuning the weights in such "autoencoder" networks, but this works well only if the initial weights are close to a good solution. We describe an effective way of initializing the weights that allows deep autoencoder networks to learn low-dimensional codes that work much better than principal components analysis as a tool to reduce the dimensionality of data.
Revisiting denoising auto-encores
  • L Gonzalo
  • S Giraldo
Symmetric ciphers based on two-dimensional chaotic maps
  • J Fridrich
Learning structured output representation using deep conditional generative models
  • K Sohn
  • H Lee
  • X Yan