Image Steganography Technique Research Articles

Securing AI Models Against Backdoor Attacks: A Novel Approach Using Image Steganography

<p>Artificial Intelligence (AI) has become ubiquitous, transforming numerous domains including traffic sign recognition, defect detection, and healthcare. However, this widespread adoption has brought about significant cybersecurity challenges, particularly in the form of backdoor attacks, which manipulate training datasets to compromise model integrity. While the integration of AI has proven beneficial, there is a lack of comprehensive strategies to protect AI models from these covert attacks, necessitating innovative approaches for securing AI systems. In this study, we demonstrate a novel methodology that integrates image steganography with deep learning techniques, aiming to obscure backdoor triggers and enhance the resilience of AI models against these attacks. We employ a diverse set of AI models and conduct extensive evaluations in a traffic sign recognition scenario, specifically targeting the STOP sign. The results reveal that shallow models are challenged in learning trigger information and are sensitive to trigger settings, while deeper models achieve an impressive 98.03% attack success rate. The image steganography technique used requires minimal data adjustments, making the triggers more challenging to detect than with traditional methods. Our findings underscore the stealth and severity of backdoor attacks, emphasizing the need for advanced security measures in AI and contributing to the broader understanding and development of robust protections against such attacks.</p> <p> </p>

GhostEncoder: Stealthy backdoor attacks with dynamic triggers to pre-trained encoders in self-supervised learning

Within the realm of computer vision, self-supervised learning (SSL) pertains to training pre-trained image encoders utilizing a substantial quantity of unlabeled images. Pre-trained image encoders can serve as feature extractors, facilitating the construction of downstream classifiers for various tasks. However, the use of SSL has led to an increase in security research related to various backdoor attacks. Currently, the trigger patterns used in backdoor attacks on SSL are mostly visible or static (sample-agnostic), making backdoors less covert and significantly affecting the attack performance. In this work, we propose GhostEncoder, the first dynamic invisible backdoor attack on SSL. Unlike existing backdoor attacks on SSL, which use visible or static trigger patterns, GhostEncoder utilizes image steganography techniques to encode hidden information into benign images and generate backdoor samples. We then fine-tune the pre-trained image encoder on a manipulation dataset to inject the backdoor, enabling downstream classifiers built upon the backdoored encoder to inherit the backdoor behavior for target downstream tasks. We evaluate GhostEncoder on three downstream tasks and results demonstrate that GhostEncoder provides practical stealthiness on images and deceives the victim model with a high attack success rate without compromising its utility. Furthermore, GhostEncoder withstands state-of-the-art defenses, including STRIP, STRIP-Cl, and SSL-Cleanse.

Enhancing Security through Advanced Image Steganography Techniques

The project aims to explore various types of steganography techniques, with a focus on image steganography, where information is concealed within images. This involves decrypting the hidden data to retrieve the message image, a process that can be achieved through different methods. The study delves into image steganography, which involves hiding information such as text, images, or audio files within other image or video files. Specifically, the project concentrates on utilizing steganography to embed an image within another image using spatial domain techniques. By examining this particular method, the project seeks to deepen understanding of how image steganography operates and its potential applications. This research contributes to the field by providing insights into the implementation and effectiveness of steganography techniques, offering valuable knowledge for information security and data protection purposes. Through practical demonstrations and analysis, the project aims to showcase the capabilities and limitations of image steganography, facilitating further research and development in this domain. Ultimately, the project endeavors to enhance comprehension of steganography methods and their significance in safeguarding digital information

Parallel BFS through pennant data structure with reducer hyper‐object based data hiding for 3D mesh images

AbstractData hiding refers to the technique employed to conceal sensitive data within various forms of multimedia, such as video, audio, text, 2D images, 3D images, and more, while also being able to successfully extract said information from the multimedia file. Although numerous data hiding techniques exist for 2D images, research on data hiding in 3D images is still in its early stages. Current 3D image steganography methods suffer from limitations in terms of embedding capacity and time complexity. To overcome these difficulties, we propose a novel 3D image steganographic technique using Parallel Breadth First Search (PBFS) with hyper‐object. Our approach utilizes the PBFS searching strategy, along with layer synchronization, to embed private information within the vertices of the 3D mesh images. To maximize the effectiveness of the process and minimize time, cost, complexity, and execution, we involved the data structure “bag” to parallelize the BFS. The implementation of bags is based on the pennant data structure. This methodology obtained an impressive Embedding Capacity (EC) of 9.00 bits per vertex (bpv) while maintaining superior visual quality and time complexity. Consequently, the proposed methodology holds great potential for widespread utilization across numerous sectors, including private and government organizations such as intelligence agencies, intellectual property rights management, cloud data security, defense, covert communication, medical imagery, and more.

Hybrid Edge Detection Methods in Image Steganography for High Embedding Capacity

Abstract In this research, we propose two new image steganography techniques focusing on increasing image-embedding capacity. The methods will encrypt and hide secret information in the edge area. We utilized two hybrid methods for the edge detection of the images. The first method combines the Laplacian of Gaussian (LoG) with the wavelet transform algorithm and the second method mixes the LOG and Canny. The Combining was performed using addWeighted. The text message will be encrypted using the GIFT cipher method for further security and low computation. For the effectiveness evaluation of the proposed method, various evaluation metrics were used such as embedding capacity, PSNR, MSE, and SSIM. The obtained results indicate that the proposed method has a greater embedding capacity in comparison with other methods, while still maintaining high levels of imperceptibility in the cover image.

A novel technique for image steganography based on maximum energy seam

AbstractImage steganography is the art of hiding information in a cover image in such a way that a third party does not notice the hidden information. This paper presents a novel technique for image steganography in the spatial domain. The new method hides and recovers hidden information of substantial length within digital imagery while maintaining the size and quality of the original image. The image gradient is used to generate a saliency image, which represents the energy of each pixel in the image. Pixels with higher energy are more salient and they are valuable for hiding data since their visual impairment is low. From the saliency image, a cumulative maximum energy matrix is created; this matrix is used to generate horizontal seams that pass over the maximum energy path. By embedding the secret bits of information along the seams, a stego-image is created which contains the hidden message. In the stego-image, we ensure that the hidden data is invisible, with very small perceived image quality degradation. The same algorithms are used to reconstruct the hidden message from the stego-image. Experiments have been conducted using two types of images and two types of hidden data to evaluate the proposed technique. The experimental results show that the proposed algorithm has a high capacity and good invisibility, with a Peak Signal-to-Noise Ratio (PSNR) of about 70, and a Structural SIMilarity index (SSIM) of about 1.

Efficient color image steganography based on new adapted chaotic dynamical system with discrete orthogonal moment transforms

In today's rapidly evolving digital landscape, safeguarding the secure transmission of sensitive information is paramount. Traditional steganography methods, while effective at data concealment, have grown vulnerable to brute force attacks in our age of powerful computing. This paper addresses the pressing need to bolster data transmission security. To address the challenges posed by limited key space and security concerns associated with 1-D chaotic maps in data transmission, we introduce a novel 1-D chaotic map, the Multi-Parameter Chebyshev Map (MCM). The MCM boasts four control parameters, each with a broad chaotic range and high chaos level, significantly enhancing security. Furthermore, we present an innovative steganography technique for color images that leverages the MCM in conjunction with discrete moment transforms. This method chaotically embeds the secret message within the moment coefficients of a cover color image using MCM and quantization rules. The expansive chaotic range and high chaos level of the MCM substantially reinforce the security of this approach. The considered moment transforms include Tchebichef, Krawtchouk, Hahn, Charlier, Meixner, Shmaliy, Racah, and Dual-Hahn. The performance of the proposed steganography method is thoroughly evaluated and compared with other state-of-the-art methods in terms of hiding capacity and image quality in both spatial and transform domains. The quantization of moment coefficients demonstrates its superior performance. This method holds crucial implications for practical applications requiring secure communication of color images, such as military and medical contexts. The combination of MCM and moment transforms in this method presents a promising avenue for future research in the field of information hiding.

Image Steganography:

Steganography is the practice of hiding secret information within other media, such as images, audio, video, and text. It has become increasingly important in today's society as a way to enable private and secure communication. This research project focuses on image steganography techniques that are used to evade detection of the secret message through statistical steganalysis techniques. The aim of this research is to compare and evaluate different image steganography methods, study their implementation complexity, and propose a framework to improve current approaches. The research will provide a comparison of the efficiency of different steganography techniques in avoiding detection by steganalysis and may lead to the development of better steganography techniques in the future. This paper focuses on the three steganography methods in the spatial domain: Least Significant Bit (LSB), Pixel-Value Difference (PVD) and Edge-based Data Embedding (EBE) methods. A simple experiment has been conducted to encrypt several images using these three methods, and the distortion measure for LSB using mean square error (MSE) and peak-to-noise ratio (PSNR) has been investigated. Although the distortion measure result is considered acceptable for LSB method in the experiment, all methods resulted in significant difference in the file capacity. This shows that further enhancement in the security of the encryption is necessary so that the secret message will not be easily discovered. Thus, in this paper, we propose a conceptualized enhancement in the security using Morse Code, Base 64, SHA-245, and Advanced Encryption Standard (AES) before encrypting using PVD.

FaceSigns: Semi-Fragile Watermarks for Media Authentication

Manipulated media is becoming a prominent threat due to the recent advances in realistic image and video synthesis techniques. There have been several attempts at detecting synthetically tampered media using machine learning classifiers. However, such classifiers do not generalize well to black-box image synthesis techniques and have been shown to be vulnerable to adversarial examples. To address these challenges, we introduce FaceSigns — a deep learning based semi-fragile watermarking technique that allows media authentication by verifying an invisible secret message embedded in the image pixels. Instead of identifying and detecting manipulated media using visual artifacts, we propose to proactively embed a semi-fragile watermark into a real image or video so that we can prove its authenticity when needed. FaceSigns is designed to be fragile to malicious manipulations or tampering while being robust to benign operations such as image/video compression, scaling, saturation, contrast adjustments etc. This allows images and videos shared over the internet to retain the verifiable watermark as long as a malicious modification technique is not applied. We demonstrate that our framework can embed a 128 bit secret as an imperceptible image watermark that can be recovered with a high bit recovery accuracy at several compression levels, while being non-recoverable when unseen malicious manipulations are applied. For a set of unseen benign and malicious manipulations studied in our work, our framework can reliably detect manipulated content with an AUC score of 0.996 which is significantly higher than prior image watermarking and steganography techniques.

Image Steganography: A Review

Abstract: An image steganographic scheme is one kind of steganographic systems, where the secret message is hidden in a digital image with some hiding method. Someone can then use a proper embedding procedure to recover the hidden message from the image. The original image is called a cover image in steganography, and the message-embedded image is called a stego image. This Steganography technique is more popular in recent year than other steganography possibly because of the flood of electronic image information available with the advent of digital cameras and high-speed internet distribution. It can involve hiding information in the naturally occurred noise within the image. In this paper we provide a review on various image steganography techniques.

Optimal location using genetic algorithms for chaotic image steganography technique based on discrete framelet transform

The development of network technology and its use have led society into the internet world and communication through digital data. The ease of access to digital resources and the possibility of unauthorized copies and distribution of them have gave rise to the significant challenge of protecting the ownership of these data. In this context, steganography is cited as a possible solution to some of these problems. Steganography can be understood as the practice of hiding secret information within another object, as for instance a text, image or video, such that the presence of this hidden information is not evident and it can be extracted if necessary. In this paper we provide an image steganography algorithm in order to insert a grayscale secret image into a RGB color cover image. This algorithm is based on a general hybrid chaos map and a 3D shift function applied to the pixels of a color image. In order to select suitable hiding places in the host image, we develop a genetic algorithm with permutation encoding aiming to optimize this task. We demonstrate empirically the effectiveness of the proposed method from various perspectives, including transparency, security, and resistance.

A Password and Least Significant Bit Substitution Based Steganography Technique for Image Hiding

The proposed work introduces an image steganography technique using text-based password that acts as the symmetric key of encryption and decryption. It contains two components, an encoder, and a decoder. An encoder takes a secret image that needs to be hidden inside another cover image. The Combined image is encrypted with an eight characters’ password. The starting position of the cover image from where the secret image starts to hide depends on the password. Then the (Least Significant Bit) LSB Substitution technique is used to hide the entire secret image inside the cover image. The decoder takes the encrypted image and the password to retrieve the secret image back. The proposed work produces an encrypted image by keeping 99% information of the cover image and 100% information of the secret image, thus it provides lossless retrieval of the secret image. The proposed work includes dynamicity of position in steganography that ensures more security of image data over the network than the existing state of art methods.

Implementing Image Steganography Techniques for Secure Data Hiding in the Development of an Android Application

Implementing Image Steganography Techniques for Secure Data Hiding in the Development of an Android Application

A Huffman code LSB based image steganography technique using multi-level encryption and achromatic component of an image

In the recent couple of years, due to the accelerated popularity of the internet, various organizations such as government offices, military, private companies, etc. use different transferring methods for exchanging their information. The Internet has various benefits and some demerits, but the primary bad mark is security of information transmission over an unreliable network, and widely uses of images. So, Steganography is the state of the art of implanting a message in the cover objects, that nobody can suspect or identify it. Therefore, in the field of cover steganography, it is very critical to track down a mechanism for concealing data by utilizing different blends of compression strategies. Amplifying the payload limit, and robustness, and working on the visual quality are the vital factors of this research to make a reliable mechanism. Different cover steganography research strategies have been recommended, and each adores its benefits and impediments but there is a need to foster some better cover steganography implements to accomplish dependability between the essential model of cover steganography. To handle these issues, in this paper we proposed a method in view of Huffman code, Least Significant Bits (LSB) based cover steganography utilizing Multi-Level Encryption (MLE) and colorless part (HC-LSBIS-MLE-AC) of the picture. It also used different substitution and flicking concepts, MLE, Magic matrix, and achromatic concepts for proving the proficiency, and significance of the method. The algorithm was also statistically investigated based on some Statistical Assessment Metrics (SAM) such as Mean Square Error (MSE), Peak Signal Noise Ratio (PSNR), Normalized Cross Correlation (NCC), Structural Similarity Index Metric (SSIM), etc. and different perspectives. The observational outcomes show the likelihood of the proposed algorithm and the capacity to give unwavering quality between security, payload, perception, computation, and temper protection.

A Blind Image Steganography Algorithm Based on Knight Tour Algorithm and QR Codes

Internet proliferation and technological progress have made multimedia information quickly accessible, but they have also posed a threat to privacy and security. Researchers have been interested in digital images due to their capacity to store large amounts of data due to the possibility of protecting sensitive information through digital steganography. Despite their visual imperceptibility, robustness, and ability to embed information, existing image steganography techniques face several challenges. To overcome these challenges, a novel image steganography approach based on a blind model strategy has been proposed for hiding covert messages. The model consists of two stages: embedding and extracting. In the embedding stage, a suitable cover image is selected using the FAST feature point detector. A text message is then converted to a QR code and embedded in the feature points' neighbors using knight tour steps in a chess game. The result is a stego image that appears identical to the original cover image but contains the secret message in its feature points' neighbors. The extracting stage involves finding the feature points and extracting the QR code to obtain the original text message. Since the feature points were not altered during the embedding process, the proposed model is known as a blind model. This approach eliminates the need for the original cover image during the extracting stage. The proposed model was evaluated using several metrics, including the peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM). The results demonstrate that the proposed algorithm can effectively embed and extract secret messages with high accuracy while maintaining the visual quality of the cover image with 100% of (SSIM), and 73.48 as an average of (PSNR).

IDGAN: Information-Driven Generative Adversarial Network of Coverless Image Steganography

Traditional image steganography techniques complete the steganography process by embedding secret information into cover images, but steganalysis tools can easily detect detectable pixel changes that lead to the leakage of confidential information. The use of a generative adversarial network (GAN) makes it possible to embed information using a combination of information and noise in generating images to achieve steganography. However, this approach is usually accompanied by issues such as poor image quality and low steganography capacity. To address these challenges, we propose a steganography model based on a novel information-driven generative adversarial network (IDGAN), which fuses a GAN, attention mechanisms, and image interpolation techniques. We introduced an attention mechanism on top of the original GAN model to improve image accuracy. In the generation model, we replaced some transposed convolution operations with image interpolation for better quality of dense images. In contrast to traditional steganographic methods, the IDGAN generates images containing confidential information without using cover images and utilizes GANs for information embedding, thus having better anti-detection capability. Moreover, the IDGAN uses an attention mechanism to improve the image details and clarity and optimizes the steganography effect through an image interpolation algorithm. Experimental results demonstrate that the IDGAN achieves an accuracy of 99.4%, 95.4%, 93.2%, and 100% on the MNIST, Intel Image Classification, Flowers, and Face datasets, respectively, with an embedding rate of 0.17 bpp. The model effectively protects confidential information while maintaining high image quality.

Combined image encryption and steganography technique for enhanced security using multiple chaotic maps

This paper proposes a novel security algorithm which combines the concepts of Cryptography and Steganography to provide enhanced security to digital images. The proposed algorithm, named CIEST, enciphers the secret image (cryptography) which is hidden into the cover image (steganography) for secured transfer of information. A two-stage encryption process, with one of the stages being DNA encoding is used here. The cover image is also modified before the data embedding phase to improve the effectiveness of the algorithm. Multiple Combined Chaotic maps are utilised in the encryption and embedding stages to improve the randomness of the algorithm. The use of chaotic maps also makes the algorithm more robust to different types of attacks. The CIEST scheme is shown to be robust, offering high embedding capacity with high sensitivity with a near zero correlation value, information entropy value of 7.9972, maximum SSIM of 0.9992, and a maximum PSNR of 56.1071. The experimental results presented establish the immunity of the proposed method against attacks along with the feasibility and effectiveness of the algorithm in real time.

Image steganography technique based on bald eagle search optimal pixel selection with chaotic encryption

In the digital era, information security becomes a challenging process that can be mitigated by the utilization of cryptography and steganography techniques. Earlier studies on steganography have the risk of exposing confidential data by an anonymous user. For resolving, the limitations related to the existing algorithms, one of the efficient solutions in encryption-based steganography. Encryption techniques act as an important part in protect actual data from illegal access. This study focuses on the design of Bald Eagle Search Optimal Pixel Selection with Chaotic Encryption (BESOPS-CE) based image steganography technique. The presented BESOPS-CE technique effectively hides the secret image in its encrypted version to the cover image. For accomplishing this, the BESOPS-CE technique employs a BES for optimal pixel selection (OPS) procedure. Besides, chaotic encryption was executed for encrypting the secret image, which is then embedded to choose pixel points of the cover image. Finally, embedding and extraction processes are carried out. The inclusion of the encryption process aids in accomplishing an added layer of security. A comprehensive simulation study was used to report on the BESOPS-CE approach's increased performance, and the results are examined from many angles. A thorough comparative analysis revealed that the BESOPS-CE model outperformed more contemporary methods.

A review on image steganographic techniques based on optimization algorithms for secret communication

A review on image steganographic techniques based on optimization algorithms for secret communication

Huffman Encoding with White Tailed Eagle Algorithm-based Image Steganography Technique

Since the Internet is a medium for transporting sensitive data, the privacy of the message transported has become a major concern. Image steganography has become a prominent tool for hiding data to ensure privacy during transfer. An efficient steganography system is essential to accomplish the best embedding capacity and maintain the other parameters at a satisfying level. Image encryption systems provide a secure and flexible system to maintain the privacy of image conversion and storage in the transmission system. Many existing image steganography methods can be attacked by various techniques, or do not support many image formats for embedding. To resolve these shortcomings, this study presents the Huffman Encoding with White Tailed Eagle Algorithm-based Image Steganography (HEWTEA-IS) technique, aiming to achieve secrecy with no compromise in image quality. The HEWTEA-IS method uses Discrete Wavelet Transform (DWT) for the decomposition of images into different subbands, and Huffman encoding to determine the embedding bits on the decomposed blocks and offer an additional layer of security. Moreover, the WTEA resolves the problem of imperceptibility by identifying the optimal probable position in the cover image for embedding secret bits. The proposed algorithm was simulated and examined in terms of different measures, and an extensive experimental analysis ensured that it is superior to other methods in several aspects.