Steganalysis Model Research Articles

An Image Steganalysis Algorithm Based on Multi-Resolution Feature Fusion

The misuse of image steganography poses significant risks to societal security. Whether images include concealed data is a critical problem of information security. Traditional convolutional layers often fail to adequately capture the global correlation of steganographic features as network depth increases, leading to redundant model parameters and missing key features, thereby weakening steganographic signal detection. To solve the problems of highly covert steganographic algorithms and the weakness of traditional methods, a steganography detection solution using multi-resolution feature fusion is presented. This approach uses a multi-resolution network to increase the interactivity from higher to lower resolution. The results of the experiments confirm that the proposed algorithm allows for a maximum accuracy of 90.56% for the embedding rate of 0.4bpp. The overall results prove that the proposed model achieves higher accuracy and better performance than some leading steganalysis models available when applied to different steganographic algorithms and embedding rate conditions.

Using Support Vector Machine to Detect Data Hiding in Color Images

Background: A steganography investigation model that uses improved dim scale List of capabilities, for identifying information inside uncompressed RGB pictures. Materials and Methods: in this paper, a bunch of 2000 RGB pictures was made utilizing normal tones, in BMP design with a size of 512 x 512 pixels. Clean pictures are Inserted secret picture information, utilizing two Payload plans, 2 pieces for every channel (BPC) and 4 pieces for each channel. The arrangement of chosen highlights comprises of 24 elements for every variety channel, and 72 highlights for each variety channel The picture, incorporates dim-level co-event grid highlights, entropy Elements, and factual proportions of variety. The list of capabilities components is determined as Individual channels, joined into picture vector highlights. The cycle depends on steganography examination On AI, utilizing Paired vector machine (SVM) support Execute the exercise manual in MATLAB. Results: The outcomes showed that the new method had overcome the benchmark with more than 80% identification accuracy Conclusion : a new Steganalysis model was introduced by using SVM. The model can recognize the presence of stowed-away information inside RGB variety pictures, removing the highlights from datasets of perfect and messy pictures and ordering utilizing the support Vector Machine calculation. the results were compared with the DNA algorithm.

A novel method for linguistic steganography by English translation using attention mechanism and probability distribution theory.

To enhance our ability to model long-range semantical dependencies, we introduce a novel approach for linguistic steganography through English translation. This method leverages attention mechanisms and probability distribution theory, known as NMT-stega (Neural Machine Translation-steganography). Specifically, to optimize translation accuracy and make full use of valuable source text information, we employ an attention-based NMT model as our translation technique. To address potential issues related to the degradation of text quality due to secret information embedding, we have devised a dynamic word pick policy based on probability variance. This policy adaptively constructs an alternative set and dynamically adjusts embedding capacity at each time step, guided by variance thresholds. Additionally, we have incorporated prior knowledge into the model by introducing a hyper-parameter that balances the contributions of the source and target text when predicting the embedded words. Extensive ablation experiments and comparative analyses, conducted on a large-scale Chinese-English corpus, validate the effectiveness of the proposed method across several critical aspects, including embedding rate, text quality, anti-steganography, and semantical distance. Notably, our numerical results demonstrate that the NMT-stega method outperforms alternative approaches in anti-steganography tasks, achieving the highest scores in two steganalysis models, NFZ-WDA (with score of 53) and LS-CNN (with score of 56.4). This underscores the superiority of NMT-stega in the anti-steganography attack task. Furthermore, even when generating longer sentences, with average lengths reaching 47 words, our method maintains strong semantical relationships, as evidenced by a semantic distance of 87.916. Moreover, we evaluate the proposed method using two metrics, Bilingual Evaluation Understudy and Perplexity, and achieve impressive scores of 42.103 and 23.592, respectively, highlighting its exceptional performance in the machine translation task.

FPFnet: Image steganalysis model based on adaptive residual extraction and feature pyramid fusion

FPFnet: Image steganalysis model based on adaptive residual extraction and feature pyramid fusion

Toward high imperceptibility deep JPEG steganography based on sparse adversarial attack

JPEG steganography is an important branch of information hiding. However, with the development of deep learning-based steganalysis models, steganography is facing great challenges. To better resist these steganalysis models, this paper proposes a deep JPEG steganography framework based on sparse adversarial attacks. According to the vulnerability of deep steganalysis models to adversarial attacks, sparse adversarial attacks are introduced into the deep JPEG steganographic structure to improve the security of stego images. In addition, to resist steganalysis from JPEG and spatial domains, both JPEG and spatial domain steganalysis models are involved in adversarial training. Finally, to further enhance the imperceptibility of adversarial stego images, the visual perception loss is designed from the perspective of human eyes. Experimental results indicate that compared with the existing methods, the proposed method has higher imperceptibility and security and can resist modern deep steganalysis models in both JPEG and spatial domains to a certain extent. The source code is available at https://github.com/imagecbj/SAE-JS.

Lightweight image steganalysis with block-wise pruning

Image steganalysis is the task of detecting a secret message hidden in an image. Deep steganalysis using end-to-end deep learning has been successful in recent years, but previous studies focused on improving detection performance rather than designing a lightweight model for practical applications. This caused a deep steganalysis model to be heavy and computationally costly, making the model infeasible to deploy in real-world applications. To address this issue, we study an effective model design strategy for lightweight image steganalysis. Considering the domain-specific characteristics of steganalysis, we propose a simple yet effective block removal strategy that progressively removes a sequence of blocks from deep classification networks. This method involves the gradual removal of convolutional neural network blocks, starting from deeper ones. By doing so, the number of parameters and FLOPs are decreased without compromising the detection performance. Experimental results show that our removal strategy makes the EfficientNet-B0 variants 9.58 times smaller and has 2.16 times fewer FLOPs than the baseline while retaining detection accuracy of 90.73% and 82.40% that are on par with the baseline on BOSSBase and ALASKA#2 datasets, respectively. Backed by our in-depth analyses, the results indicate that only a few early layers are sufficient for effective image steganalysis.

Multi-resolution network based image steganalysis model

Multi-resolution network based image steganalysis model

FedSpy: A Secure Collaborative Speech Steganalysis Framework Based on Federated Learning

Deep learning brings the opportunity to achieve effective speech steganalysis in speech signals. However, the speech samples used to train speech steganalysis models (i.e., steganalyzers) are usually sensitive and distributed among different agencies, making it impractical to train an effective centralized steganalyzer. Therefore, in this paper, we present an effective framework, named FedSpy, using federated learning, which enables multiple agencies to securely and jointly train the speech steganalysis models without sharing their speech samples. FedSpy is a flexible and extensible framework that can work effectively in conjunction with various deep-learning-based speech steganalysis methods. We evaluate the performance of FedSpy by detecting the most widely used Quantization Index Modulation-based speech steganography with three state-of-the-art deep-learning-based steganalysis methods representatively. The results show that FedSpy significantly outperforms the local steganalyzers and achieves good detection accuracy comparable to the centralized steganalyzer.

Enhancing image steganography via adversarial optimization of the stego distribution

Adversarial steganography aims to fool steganalysis models and enhance embedding security. Its application in the sequential min-max steganographic game has recently shown state-of-the-art embedding security. However, current methods suffer from high computational costs and inferior convergence. These issues could be caused by unnecessary optimization on distortions and the monotonous criterion for adversaries. This paper introduces a novel and concise scheme for conducting adversarial steganography and refines the criterion for adversaries. The suggested adversarial optimizations leave out distortions and focus on optimizing the stego distribution, whose difference with the intrinsic cover distribution is the primary factor affecting embedding security. These optimization techniques are accomplished by utilizing gradient descent on embedding modification probabilities and introducing a novel protocol for cover enhancement. Moreover, the proposed criteria for adversaries prioritize enhancing the similarity between each cover-stego pair, rather than evading steganalysis models, to break monotony and enhance convergence. The discussion and refinement of the loss function and termination setting of the optimization process are also provided. The proposed method demonstrates its benefits in increased security and computation efficiency in experiments. After collaborating with the steganographic min-max game, the state-of-the-art embedding security and stronger resistance to various and unexpected steganalysis models are seen.

STFF-SM: Steganalysis Model Based on Spatial and Temporal Feature Fusion for Speech Streams

The real-time detection of speech steganography in Voice-over-Internet-Protocol (VoIP) scenarios remains an open problem, as it requires steganalysis methods to perform for low-intensity embeddings and short-sample inputs, as well as provide rapid detection results. To address these challenges, this paper presents a novel steganalysis model based on spatial and temporal feature fusion (STFF-SM). Differing from the existing methods, we take both the integer and fractional pitch delays as input, and design subframe-stitch module to organically integrate subframe-wise integer delays and frame-wise fractional pitch delays. Further, we design a spatial fusion module based on pre-activation residual convolution to extract the pitch spatial features and gradually increase their dimensions to discover finer steganographic distortions to enhance the detection effect, where a Group-Squeeze-Weighting block is introduced to alleviate the information loss in the process of increasing the feature dimension. In addition, we design a temporal fusion module to extract pitch temporal features using the stacked LSTM, where a Gated Feed-Forward Network is introduced to learn the interaction between different feature maps while suppressing the features that are not useful for detection. We evaluated the performance of STFF-SM through comprehensive experiments and comparisons with the state-of-the-art solutions. The experimental results demonstrate that STFF-SM can well meet the needs of real-time detection of speech steganography in VoIP streams, and outperforms the existing methods in detection performance, especially with low embedding strengths and short window sizes.

STD-NET: Search of Image Steganalytic Deep-Learning Architecture Via Hierarchical Tensor Decomposition

Steganalysis aims to reveal covert communication established via steganography. In the arm race with steganography, steganalysis has evolved from the old-style hand-crafted features set to deep-learning architectures. However, recent studies show that the majority of existing deep steganalysis models have a large amount of redundancy, which leads to a huge waste of storage and computing resources. The existing model compression method cannot flexibly compress the convolutional layer in residual shortcut block so that a satisfactory shrinking rate cannot be obtained. In this paper, we propose STD-NET, an unsupervised deep-learning architecture search approach via hierarchical tensor decomposition for image steganalysis. Our proposed strategy will not be restricted by various residual connections, since this strategy does not change the number of input and output channels of the convolution block. We propose a normalized distortion threshold to evaluate the sensitivity of each involved convolutional layer of the base model to guide STD-NET to compress target network in an efficient and unsupervised approach, and obtain two network structures of different shapes with low computation cost and similar performance compared with the original one. Extensive experiments have confirmed that, on one hand, our model can achieve comparable or even better detection performance in various steganalytic scenarios due to the great adaptivity of the obtained network architecture. On the other hand, the experimental results also demonstrate that our proposed strategy is more efficient and can remove more redundancy compared with previous steganalytic network compression methods.

Linguistic Steganalysis by Enhancing and Integrating Local and Global Features

With the improvement of steganography, the difference of statistical distribution caused by information hiding is getting smaller and smaller, increasing the difficulty of steganalysis. The existing steganalysis models value different features equally, while the differences in the importance of high-dimensional features are ignored. That is the influences of feature quality on model performance are not considered. To fill the gap, a novel linguistic steganalysis by enhancing and integrating local and global features is proposed in this letter. It extracts and integrates the features of two dimensions, namely local semantic features and global long-term dependencies, to construct a joint feature map. Then, to improve the quality of features, a group-wise enhancement mechanism is employed, which divides features into multiple groups, enhancing important features in each group while weakening the less important ones by generating an important coefficient for each sub-feature in each semantic group. Finally, the enhanced features are used to further extract high-quality text representation to help the classification module distinguish cover and stego texts more accurately. The experimental results show that the proposed model has better detection performance in multiple hidden scenarios. The ablation experiment manifests that using integrated and enhanced features to extract high-quality text representations can effectively improve the discrimination ability of the model.

Extensible Steganalysis via Continual Learning

To realize secure communication, steganography is usually implemented by embedding secret information into an image selected from a natural image dataset, in which the fractal images have occupied a considerable proportion. To detect those stego-images generated by existing steganographic algorithms, recent steganalysis models usually train a Convolutional Neural Network (CNN) on the dataset consisting of paired cover/stego-images. However, it is inefficient and impractical for those steganalysis models to completely retrain the CNN model to make it effective for detecting a new emerging steganographic algorithm while maintaining the ability to detect the existing steganographic algorithms. Thus, those steganalysis models usually lack dynamic extensibility for new steganographic algorithms, which limits their application in real-world scenarios. To address this issue, we propose an accurate parameter importance estimation (APIE)-based continual learning scheme for steganalysis. In this scheme, when a steganalysis model is trained on a new image dataset generated by a new emerging steganographic algorithm, its network parameters are effectively and efficiently updated with sufficient consideration of their importance evaluated in the previous training process. This scheme can guide the steganalysis model to learn the patterns of the new steganographic algorithm without significantly degrading the detectability against the previous steganographic algorithms. Experimental results demonstrate the proposed scheme has promising extensibility for new emerging steganographic algorithms.

Highly accurate end-to-end image steganalysis based on auxiliary information and attention mechanism

The purpose of image steganalysis is to detect whether the secret information is hidden in the image. The current advanced adaptive steganography algorithm hides the secret information in the complex area of the image texture. However, it is difficult for the image steganalysis model to capture enough noise residual, which makes the detection ability of the model insufficient. To further improve the detection ability of the spatial image steganalysis model, a U-Net-based auxiliary information generation network is first constructed to increase the size of noise residual in the image so that the model can capture more favorable information. Besides, the spatial and channel attention mechanisms are fused to guide the model to pay more attention to the regions of the image that are globally favorable for steganalysis. To verify the effectiveness of the proposed model, experiments are conducted on the BOSSbase-v1.01 dataset through advanced spatial adaptive steganography algorithms S-UNIWARD and WOW. The experimental results show that the proposed model can improve the detection accuracy by up to 4.5% compared with the current optimal deep learning-based spatial image steganalysis models SR-Net and Siamese-Net.

Preprocessing Enhancement Method for Spatial Domain Steganalysis

In the field of steganalysis, in recent years, the research focus has mostly been on optimizing the structures of neural networks, while the application of high-pass filters is still limited to the simple selection of filters and simple adjustment of the number of filters. In this paper, we propose a method to enhance the assistance and contribution of high-pass filters to the detection capability of a spatial domain steganalysis model, which mainly contains the preprocessing enhancement of high-pass filters and cross-layer enhancement of high-pass filters, and we construct a preprocessing enhancement model, the HPF-Enhanced Model, for spatial domain steganalysis, based on Yedroudj-Net. In the experimental part, we find the best preprocessing enhancement method through various validations, and we compare the HPF-Enhanced Model with the classical models. The results show that the proposed enhancement method can bring a significant improvement, and they also show that the preprocessing enhancement method can help to reduce the model size, and it thus can be used to construct a lightweight spatial domain steganalysis model with strong performance.

Binary Image Steganalysis Based on Symmetrical Local Residual Patterns

Residual computation is an effective method for gray-scale image steganalysis. For binary images, the residual computation calculated by the XOR operation is also employed in the local residual patterns (LRP) model for steganalysis. A binary image steganalytic scheme based on symmetrical local residual patterns (SLRP) is proposed. The symmetrical relationships among residual patterns are introduced that make the features more compact while reducing the dimensionality of the features set. Multi-scale windows are utilized to construct three SLRP submodels which are further merged to construct the final features set instead of a single model. SLRPs with higher probability to be modified after embedding are emphasized and selected to construct the feature sets for training the support vector machine classifier. The experimental results show that the proposed steganalytic scheme is effective for detecting binary image steganography.

Steganalysis of adaptive multi-rate speech streams with distributed representations of codewords

With the wide application of adaptive multi-rate (AMR) speech coder, steganography and steganalysis based on AMR coded speech streams have become a hot spot in the field of information hiding. Quantization Index Modulation (QIM)-based steganography is one of the most effective approaches to hide secret information into AMR coded speech streams with excellent imperceptibility and robustness. So far, accurate detection for QIM-based steganography in short-length or low-embedding-rate speech streams remains an open problem, though some approaches can complete detection in a short time. To address this challenge, we first analyze and verify the characteristics of QIM-based steganography, and present a novel and high time-efficient steganalysis model based on distributed representations. Specifically, a codeword embedding layer is introduced to capture distributed representations with a denser space; then we introduce a bidirectional Long Short-Term Memory (LSTM) layer and propose a gated attention mechanism to provide contextual distribution features with better generalization capabilities; finally, a Multi-Layer Perceptron (MLP) classifier is designed to distinguish normal or steganographic objects. The experimental results demonstrate that the proposed model can effectively detect QIM-based steganography in AMR speech streams and outperform the state-of-the-art ones.

Lightweight AAC Audio Steganalysis Model Based on ResNeXt

Traditional AAC (Advanced Audio Coding) audio steganalysis methods rely on manual feature extraction, which results in low detection accuracy and low efficiency. Nowadays, the new steganalysis model based on neural network is very attractive, but its scale is large and its detection accuracy needs further improvement. Aiming at the above problems, this paper proposes a lightweight AAC audio general steganalysis model based on ResNeXt network. Firstly, the residual signal of QMDCT (Quantized Modified Discrete Cosine Transform) coefficients is calculated through a fixed convolution layer composed of multiple sets of high-pass filters. Then, based on the original structure of ResNeXt network, two ResNeXt blocks are designed to form a residual learning module, by which the steganalysis features in the QMDCT coefficients are further extracted. Finally, the classification module consisting of the fully connected layer and the Softmax layer is designed to obtain the classification result. The experimental results show that the model detection accuracy can reach more than 94% under all relative embedding rates when it operates on both the steganography algorithm based on the small value area of the QMDCT coefficient and the steganography algorithm based on the Huffman code sign bit. For the algorithm based on Huffman codeword mapping, even with the relative embedding rate of 0.1, the detection accuracy of the model can reach 85.5%, which is obviously better than the existing steganalysis schemes. Compared with other steganalysis schemes based on neural network, the model in this paper has fewer parameters, and reduces the scale by more than 40%, which is more lightweight and more efficient.

Dataset mismatched steganalysis using subdomain adaptation with guiding feature

The generalization problem in deep learning has always been an important problem to be solved. In the field of steganalysis, generalization is also an important factor that makes steganalysis models difficult to deploy in real-world scenarios. For a group of suspicious images that never appeared in the training set, the pre-trained deep learning-based steganalysis models tend to suffer from distinct performance degradation. To address this limitation, in this paper, a feature-guided subdomain adaptation steganalysis framework is proposed to improve the performance of the pre-trained models when detecting new data. Initially, the source domain and target domain will be divided into subdomains according to class, and the distributions of the relevant subdomains are aligned by subdomain adaptation. Afterward, the guiding feature is generated to make the division of subdomains more stable and precise. When it is used to detect three spatial steganographic algorithms with a wide variety of datasets and payloads, the experimental results show that the proposed steganalysis framework can significantly improve the average accuracy of SRNet model by 5.4% at 0.4bpp, 8.5% at 0.2bpp, and 8.0% at 0.1bpp in the case of dataset mismatch.

Software and hardware-based audio-video crypto steganalysis model for enhancing robustness and imperceptibility of secured data

Software and hardware-based audio-video crypto steganalysis model for enhancing robustness and imperceptibility of secured data