Image steganography based on digital holography and saliency map

Image watermarking plays a vital role in providing protection from copyright violation. However, conventional watermarking techniques typically exhibit trade-offs in terms of image quality, robustness and capacity constrains. More often than not, these techniques optimize on one constrain while settling with the two other constraints. Therefore, in this paper, an enhanced saliency detection based watermarking method is proposed to simultaneously improve quality, capacity, and robustness. First, the enhanced structured matrix decomposition (E-SMD) is proposed to extract salient regions in the host image for producing a saliency mask. This mask is then applied to partition the foreground and background of the host and watermark images. Subsequently, the watermark (with the same dimension of host image) is shuffled using multiple Arnold and Logistic chaotic maps, and the resulting shuffled-watermark is embedded into the wavelet domain of the host image. Furthermore, a filtering operation is put forward to estimate the original host image so that the proposed watermarking method can also operate in blind mode. In the best case scenario, we could embed a 24-bit image as the watermark into another 24-bit image while maintaining an average SSIM of 0.9999 and achieving high robustness against commonly applied watermark attacks. Furthermore, as per our best knowledge, with high payload embedding, the significant improvement in these features (in terms of saliency, PSNR, SSIM, and NC) has not been achieved by the state-of-the-art methods. Thus, the outcomes of this research realizes a trade-off independent image watermarking method, which is a first of its kind in this domain.

Presented is an efficient watermarking scheme using Particle Swarm Optimization (PSO) to watermark host fingerprint images with their corresponding facial images in the Discrete Cosine Transform (DCT) domain. PSO is used to find the best DCT coefficients’ locations in the host image where the facial image data can be embedded, so that the distortion produced in the host image is minimum. The objective function for PSO is formulated in terms of the Structural Similarity Index (SSIM) and the Orientation Certainty Level Index (OCL) so as to base it on the simple visual effect of the human visual perception capability and correct minutia prediction ability. The results exhibit better watermarked image quality while retaining the feature set of the original fingerprint. Moreover, the proposed technique is robust so that the extraction of watermark is possible even after the watermarked image is exposed to attacks. As a result, at the receiver’s end, the watermarked fingerprint image and the extracted facial image can be verified for a secure and accurate biometric based personal authentication.

Digital watermarking is an encryption technique commonly used to protect intellectual property and copyright. Although watermarks are a robust method of protecting property rights, environmental interference in image propagation through the Internet is inevitable, and human-based image modification can also destroy watermarks. In this study, watermarks (affected by noise interference) were embedded in a 256ź×ź256 pixel host image by using the discrete cosine transform (DCT) technique, which transfers the spatial domain of the host image into the frequency domain. Subsequently, 32ź×ź32 pixel watermarked images were embedded as watermark identification codes in the transferred frequency-domain-based host image. The inverse discrete cosine transform (IDCT) technique was used to alter the frequency of the spatial domain, thereby allowing the host image to be visible to the human eye. Several common interferences, such as salt-and-pepper noise, Gaussian noise, clipping, and rotation, were used to destroy the watermarked image. After the destruction process, the watermark was almost discernible in a slightly damaged state, but difficult to identify in a seriously damaged state after using the DCT watermarking scheme. In this study, a back-propagation neural network (BPNN) algorithm combined with a DCT watermarking scheme was used to suppress the interference affecting watermarks. The simulation results of using the proposed DCT-BPNN method indicated that the original watermarked image was restored considerably after being subjected to environmental interference during image propagation.

The main objectives of digital image watermarking schemes are to provide acceptable level of both imperceptibility and robustness for wide variety of host and watermark images. To support watermarking scheme in accomplishing these objectives, suitable watermark strength is identified in the form of scaling factor using Dynamic-PSO (DPSO) for colored images. A non-blind, DWT-SVD-based watermarking scheme is used to embed watermark in host image. Results achieved through DPSO are compared with PSO and other widely accepted variants of PSO. Experiment results demonstrate that performance of underline watermarking scheme when used with DPSO is better than other variants of PSO.

When host images are watermarked with CRT (Chinese Remainder Theorem), the watermark images are still robust in spite of the damage of the host images by maintaining the remainders in an unchanged state within some range of the changes that are incurred by the attacks. This advantage can also be attained by "zero-watermarking," which does not change the host images in any way. This paper proposes an improved zero-watermarking scheme for color images on the DCT (Discrete Cosine Transform) domain that is based on the CRT. In the scheme, RGB images are converted into YCbCr images, and one channel is used for the DCT transformation. A key is then computed from the DC and three low-frequency AC values of each DCT block using the CRT. The key finally becomes the watermark key after it is combined four times with a scrambled watermark image. When watermark images are extracted, each bit is determined by majority voting. This scheme shows that watermark images are robust against a number of common attacks such as sharpening, blurring, JPEG lossy compression, and cropping.

Several methods are exploited to watermark digital images as a safety measure for storing information, but an attacker can destroy the information by cropping a segment of the watermarked image. In recent years, numerous schemes were proposed that reduce the impact of such attacks. A new method has been proposed to confront cropping attack that is carried out using two sudoku tables. In this method, the watermark image is scattered in two sudoku table layouts with different solutions and is watermarked in the host image. Using this method, the watermark image is repeated 81 times in the host image, and to this effect the watermark image can be reconstructed using other segments when cropped by the attacker. Both sudokus used in this paper are in the classic 9×9 form and using this method, resistance to cropping attacks increases up to 98.8%.

In this thesis, we describe new approaches and methods for reconstructing complex-valued wave fields from digital holograms. We focus on Fresnel holograms recorded in an off-axis geometry, for which operational real-time acquisition setups readily exist. The three main research directions presented are the following. First, we derive the necessary tools to port methods and concepts of wavelet-based approaches to the field of digital holography. This is motivated by the flexibility, the robustness, and the unifying view that such multiresolution procedures have brought to many applications in image processing. In particular, we put emphasis on space-frequency processing and sparse signal representations. Second, we propose to decouple the demodulation from the propagation problem, which are both inherent to digital Fresnel holography. To this end, we derive a method for retrieving the amplitude and phase of the object wave through a local analysis of the hologram's interference fringes. Third, since digital holography reconstruction algorithms involve a number of parametric models, we propose automatic adjustment methods of the corresponding parameters. We start by investigating the Fresnel transform, which plays a central role in both the modeling of the acquisition procedure and the reconstruction of complex wave fields. The study of the properties that are central to wavelet and multiresolution analysis leads us to derive Fresnelets, a new family of wavelet-like bases. Fresnelets permit the analysis of holograms with a good localization in space and frequency, in a way similar to wavelets for images. Since the relevant information in a Fresnel off-axis hologram may be separated both in space and frequency, we propose an approach for selectively retrieving the information in the Fresnelet domain. We show that in certain situations, this approach is superior to others that exclusively rely on the separation in space or frequency. We then derive a least-squares method for the estimation of the object wave's amplitude and phase. The approach, which is reminiscent of phase-shifting techniques, is sufficiently general to be applied in a wide variety of situations, including those dictated by the use of microscopy objectives. Since it is difficult to determine the reconstruction distance manually, we propose an automatic procedure. We take advantage of our separate treatment of the phase retrieval and propagation problems to come up with an algorithm that maximizes a sharpness metric related to the sparsity of the signal's expansion in distance-dependent Fresnelet bases. Based on a simulation study, we suggest a number of guidelines for deciding which algorithm to apply to a given problem. We compare existing and the newly proposed solutions in a wide variety of situations. Our final conclusion is that the proposed methods result in flexible algorithms that are competitive with preexisting ones and superior to them in many cases. Overall, they may be applied in a wide range of experimental situations at a low computational cost.

Digital watermarking provides the solution for problems such as copyright and authentication in today’s digital world of communication. Trustworthy watermarking techniques should provide the protection for rightful ownership, robustness and resist against attacks such as filtering and modification. In this paper, we proposed an efficient watermarking scheme based on singular value decomposition (SVD) and quantum Hilbert image scrambling. In the proposed watermarking scheme, first watermark is scrambled using quantum Hilbert image scrambling then perform alpha blending on scrambled watermark and SVD transformed cover image followed by alpha mixing operation. Finally, perform SVD operation of the generated signal; after alpha blending operation, we get watermarked image. The experimental results show strength of the proposed technique in terms of the robustness and security. Performance of the proposed scheme is tested on different image quality parameters of watermarked and host image. These are peak signal-to-noise ratio (PSNR), mean square error (MSE) and normalized cross-correlation (NCC), average difference (AD), maximum difference (MD), and structural contents (SC).KeywordsSingular value decompositionWatermarkQuantum Hilbert image scramblingSteganography

Three-dimensional object recognition using joint fractional Fourier transform correlators with the help of digital Fresnel holography

An enhanced hybrid image watermarking scheme for security of medical and non-medical images based on DWT and 2-D SVD

Digital in-line holography is commonly used to reconstruct 3D images from 2D holograms for microscopic objects. One of the technical challenges that arise in the signal processing stage is removing the twin image that is caused by the phase-conjugate wavefront from the recorded holograms. Twin image removal is typically formulated as a non-linear inverse problem due to the irreversible scattering process when generating the hologram. Recently, end-to-end deep learning-based methods have been utilized to reconstruct the object wavefront (as a surrogate for the 3D structure of the object) directly from a single-shot in-line digital hologram. However, massive data pairs are required to train deep learning models for acceptable reconstruction precision. In contrast to typical image processing problems, well-curated datasets for in-line digital holography does not exist. Also, the trained model highly influenced by the morphological properties of the object and hence can vary for different applications. Therefore, data collection can be prohibitively cumbersome in practice as a major hindrance to using deep learning for digital holography. In this paper, we proposed a novel implementation of autoencoder-based deep learning architecture for single-shot hologram reconstruction solely based on the current sample without the need for massive datasets to train the model. The simulations results demonstrate the superior performance of the proposed method compared to the state of the art single-shot compressive digital in-line hologram reconstruction method.

Digital watermarking has been widely studied as a method of protecting the intellectual property rights of digital images, which are high value-added contents. Recently, studies implementing these techniques with neural networks have been conducted. This paper also proposes a neural network to perform a robust, invisible blind watermarking for digital images. It is a convolutional neural network (CNN)-based scheme that consists of pre-processing networks for both host image and watermark, a watermark embedding network, an attack simulation for training, and a watermark extraction network to extract watermark whenever necessary. It has three peculiarities for the application aspect: The first is the host image resolution’s adaptability. This is to apply the proposed method to any resolution of the host image and is performed by composing the network without using any resolution-dependent layer or component. The second peculiarity is the adaptability of the watermark information. This is to provide usability of any user-defined watermark data. It is conducted by using random binary data as the watermark and is changed each iteration during training. The last peculiarity is the controllability of the trade-off relationship between watermark invisibility and robustness against attacks, which provides applicability for different applications requiring different invisibility and robustness. For this, a strength scaling factor for watermark information is applied. Besides, it has the following structural or in-training peculiarities. First, the proposed network is as simple as the most profound path consists of only 13 CNN layers, which is through the pre-processing network, embedding network, and extraction network. The second is that it maintains the host’s resolution by increasing the resolution of a watermark in the watermark pre-processing network, which is to increases the invisibility of the watermark. Also, the average pooling is used in the watermark pre-processing network to properly combine the binary value of the watermark data with the host image, and it also increases the invisibility of the watermark. Finally, as the loss function, the extractor uses mean absolute error (MAE), while the embedding network uses mean square error (MSE). Because the extracted watermark information consists of binary values, the MAE between the extracted watermark and the original one is more suitable for balanced training between the embedder and the extractor. The proposed network’s performance is confirmed through training and evaluation that the proposed method has high invisibility for the watermark (WM) and high robustness against various pixel-value change attacks and geometric attacks. Each of the three peculiarities of this scheme is shown to work well with the experimental results. Besides, it is exhibited that the proposed scheme shows good performance compared to the previous methods.

Technology as a means of conveying information can have significant implications for the security and confidentiality of data and information content. When a sender wishes to convey a message to someone and does not want the message to be easily visible, the concealment of this message becomes crucial for the sender. One commonly used method for data concealment is the Least Significant Bit (LSB) method in steganography. The LSB method is preferred for its simplicity, fast process, and greater capacity. Messages are embedded by replacing the least significant bit in the host image with the message bits, with the rule that the number of message bits does not exceed the number of bits in the host image. Based on research findings, the results of steganography are difficult to distinguish from the original content through visual observation. Therefore, it can be concluded that the steganography method is highly effective for concealing a message.

A novel DCT domain color digital watermarking scheme based on chaotic maps is proposed, and the host image is color RGB image. The color space of the color host image is first converted from RGB to YIQ. In order to resist JPEG compression, all three watermark RGB channel signals are buried to Y component of the YIQ color space host images. In the proposed method, the image watermark is embedded into low-band DCT coefficients with different strength according to the characteristics of human vision system. The watermark signal and the embedding position of the host image are encrypted respectively by the unified chaotic map and multilayer Arnold cat map, and compared with the general Arnold cat map, the key space is highly increased. The simulation results show that the proposed scheme possesses the invisibility and the robustness to common image processing attacks, such as noise addition, JPEG compression, image rotation, cutting and so on.

In this paper, a novel medical image watermarking (MIW) technique for tele ‐ medicine applications is proposed. In this approach homomorphic transform (HT), redundant discrete wavelet transform (RDWT) and singular value decomposition (SVD) transforms are successfully utilized for developing an effective watermarking technique. Reflectance component of medical host image is obtained using HT, on which RDWT decomposition and SVD are applied. Similarly medical watermark image is processed with RDWT and SVD. Singular values (Svs) were used for embedding the watermark image into host image. 2‐D chaotic Arnold transform (AT) is used for providing extra security to the watermark image by encrypting the watermarked image. Moreover, proposed technique is also tested for several wavelet families. Robustness and imperceptibility of the proposed technique is enhanced as depicted in experimental results under various attacks. Furthermore, better robustness is attained from proposed technique on comparing it with other formerly reported schemes.