arXiv subjects

Hanzhou Wu

Publications and source records attributed to Hanzhou Wu.

At least 37 records · Page 2Linked to original sources

arXiv2023

Robust and Imperceptible Black-box DNN Watermarking Based on Fourier Perturbation Analysis and Frequency Sensitivity Clustering

Recently, more and more attention has been focused on the intellectual property protection of deep neural networks (DNNs), promoting DNN watermarking to become a hot research topic. Compared with embedding watermarks directly into DNN parameters, inserting trigger-set watermarks enables us to verify the ownership without knowing the internal details of the DNN, which is more suitable for application scenarios. The cost is we have to carefully craft the trigger samples. Mainstream methods construct the trigger samples by inserting a noticeable pattern to the clean samples in the spatial domain, which does not consider sample imperceptibility, sample robustness and model robustness, and therefore has limited the watermarking performance and the model generalization. It has motivated the authors in this paper to propose a novel DNN watermarking method based on Fourier perturbation analysis and frequency sensitivity clustering. First, we analyze the perturbation impact of different frequency components of the input sample on the task functionality of the DNN by applying random perturbation. Then, by K-means clustering, we determine the frequency components that result in superior watermarking performance for crafting the trigger samples. Our experiments show that the proposed work not only maintains the performance of the DNN on its original task, but also provides better watermarking performance compared with related works.

cs.CR↗

arXiv2023

JPEG Steganalysis Based on Steganographic Feature Enhancement and Graph Attention Learning

The purpose of image steganalysis is to determine whether the carrier image contains hidden information or not. Since JEPG is the most commonly used image format over social networks, steganalysis in JPEG images is also the most urgently needed to be explored. However, in order to detect whether secret information is hidden within JEPG images, the majority of existing algorithms are designed in conjunction with the popular computer vision related networks, without considering the key characteristics appeared in image steganalysis. It is crucial that the steganographic signal, as an extremely weak signal, can be enhanced during its representation learning process. Motivated by this insight, in this paper, we introduce a novel representation learning algorithm for JPEG steganalysis that is mainly consisting of a graph attention learning module and a feature enhancement module. The graph attention learning module is designed to avoid global feature loss caused by the local feature learning of convolutional neural network and reliance on depth stacking to extend the perceptual domain. The feature enhancement module is applied to prevent the stacking of convolutional layers from weakening the steganographic information. In addition, pretraining as a way to initialize the network weights with a large-scale dataset is utilized to enhance the ability of the network to extract discriminative features. We advocate pretraining with ALASKA2 for the model trained with BOSSBase+BOWS2. The experimental results indicate that the proposed algorithm outperforms previous arts in terms of detection accuracy, which has verified the superiority and applicability of the proposed work.

cs.MM↗

arXiv2022

Robust and Lossless Fingerprinting of Deep Neural Networks via Pooled Membership Inference

Deep neural networks (DNNs) have already achieved great success in a lot of application areas and brought profound changes to our society. However, it also raises new security problems, among which how to protect the intellectual property (IP) of DNNs against infringement is one of the most important yet very challenging topics. To deal with this problem, recent studies focus on the IP protection of DNNs by applying digital watermarking, which embeds source information and/or authentication data into DNN models by tuning network parameters directly or indirectly. However, tuning network parameters inevitably distorts the DNN and therefore surely impairs the performance of the DNN model on its original task regardless of the degree of the performance degradation. It has motivated the authors in this paper to propose a novel technique called pooled membership inference (PMI) so as to protect the IP of the DNN models. The proposed PMI neither alters the network parameters of the given DNN model nor fine-tunes the DNN model with a sequence of carefully crafted trigger samples. Instead, it leaves the original DNN model unchanged, but can determine the ownership of the DNN model by inferring which mini-dataset among multiple mini-datasets was once used to train the target DNN model, which differs from previous arts and has remarkable potential in practice. Experiments also have demonstrated the superiority and applicability of this work.

cs.CR↗

arXiv2022

Generative Model Watermarking Based on Human Visual System

Intellectual property protection of deep neural networks is receiving attention from more and more researchers, and the latest research applies model watermarking to generative models for image processing. However, the existing watermarking methods designed for generative models do not take into account the effects of different channels of sample images on watermarking. As a result, the watermarking performance is still limited. To tackle this problem, in this paper, we first analyze the effects of embedding watermark information on different channels. Then, based on the characteristics of human visual system (HVS), we introduce two HVS-based generative model watermarking methods, which are realized in RGB color space and YUV color space respectively. In RGB color space, the watermark is embedded into the R and B channels based on the fact that HVS is more sensitive to G channel. In YUV color space, the watermark is embedded into the DCT domain of U and V channels based on the fact that HVS is more sensitive to brightness changes. Experimental results demonstrate the effectiveness of the proposed work, which improves the fidelity of the model to be protected and has good universality compared with previous methods.

cs.CV↗

arXiv2022

Reversible Data Hiding in Encrypted Text Using Paillier Cryptosystem

Reversible Data Hiding in Encrypted Domain (RDHED) is an innovative method that can keep cover information secret and allows the data hider to insert additional information into it. This article presents a novel data hiding technique in an encrypted text called Reversible Data Hiding in Encrypted Text (RDHET). Initially, the original text is converted into their ASCII values. After that, the Paillier cryptosystem is adopted to encrypt all ASCII values of the original text and send it to the data hider for further processing. At the data hiding phase, the secret data are embedded into homomorphically encrypted text using a technique that does not lose any information, i.e., the homomorphic properties of the Paillier cryptosystem. Finally, the embedded secret data and the original text are recovered at the receiving end without any loss. Experimental results show that the proposed scheme is vital in the context of encrypted text processing at cloud-based services. Moreover, the scheme works well, especially for the embedding phase, text recovery, and performance on different security key sizes.

cs.CR↗

arXiv2022

AWEncoder: Adversarial Watermarking Pre-trained Encoders in Contrastive Learning

As a self-supervised learning paradigm, contrastive learning has been widely used to pre-train a powerful encoder as an effective feature extractor for various downstream tasks. This process requires numerous unlabeled training data and computational resources, which makes the pre-trained encoder become valuable intellectual property of the owner. However, the lack of a priori knowledge of downstream tasks makes it non-trivial to protect the intellectual property of the pre-trained encoder by applying conventional watermarking methods. To deal with this problem, in this paper, we introduce AWEncoder, an adversarial method for watermarking the pre-trained encoder in contrastive learning. First, as an adversarial perturbation, the watermark is generated by enforcing the training samples to be marked to deviate respective location and surround a randomly selected key image in the embedding space. Then, the watermark is embedded into the pre-trained encoder by further optimizing a joint loss function. As a result, the watermarked encoder not only performs very well for downstream tasks, but also enables us to verify its ownership by analyzing the discrepancy of output provided using the encoder as the backbone under both white-box and black-box conditions. Extensive experiments demonstrate that the proposed work enjoys pretty good effectiveness and robustness on different contrastive learning algorithms and downstream tasks, which has verified the superiority and applicability of the proposed work.

cs.CV↗

arXiv2022

General Framework for Reversible Data Hiding in Texts Based on Masked Language Modeling

With the fast development of natural language processing, recent advances in information hiding focus on covertly embedding secret information into texts. These algorithms either modify a given cover text or directly generate a text containing secret information, which, however, are not reversible, meaning that the original text not carrying secret information cannot be perfectly recovered unless much side information are shared in advance. To tackle with this problem, in this paper, we propose a general framework to embed secret information into a given cover text, for which the embedded information and the original cover text can be perfectly retrieved from the marked text. The main idea of the proposed method is to use a masked language model to generate such a marked text that the cover text can be reconstructed by collecting the words of some positions and the words of the other positions can be processed to extract the secret information. Our results show that the original cover text and the secret information can be successfully embedded and extracted. Meanwhile, the marked text carrying secret information has good fluency and semantic quality, indicating that the proposed method has satisfactory security, which has been verified by experimental results. Furthermore, there is no need for the data hider and data receiver to share the language model, which significantly reduces the side information and thus has good potential in applications.

cs.CR↗

arXiv2022

Verifying Integrity of Deep Ensemble Models by Lossless Black-box Watermarking with Sensitive Samples

With the widespread use of deep neural networks (DNNs) in many areas, more and more studies focus on protecting DNN models from intellectual property (IP) infringement. Many existing methods apply digital watermarking to protect the DNN models. The majority of them either embed a watermark directly into the internal network structure/parameters or insert a zero-bit watermark by fine-tuning a model to be protected with a set of so-called trigger samples. Though these methods work very well, they were designed for individual DNN models, which cannot be directly applied to deep ensemble models (DEMs) that combine multiple DNN models to make the final decision. It motivates us to propose a novel black-box watermarking method in this paper for DEMs, which can be used for verifying the integrity of DEMs. In the proposed method, a certain number of sensitive samples are carefully selected through mimicking real-world DEM attacks and analyzing the prediction results of the sub-models of the non-attacked DEM and the attacked DEM on the carefully crafted dataset. By analyzing the prediction results of the target DEM on these carefully crafted sensitive samples, we are able to verify the integrity of the target DEM. Different from many previous methods, the proposed method does not modify the original DEM to be protected, which indicates that the proposed method is lossless. Experimental results have shown that the DEM integrity can be reliably verified even if only one sub-model was attacked, which has good potential in practice.

cs.CR↗

arXiv2022

Autoregressive Linguistic Steganography Based on BERT and Consistency Coding

Linguistic steganography (LS) conceals the presence of communication by embedding secret information into a text. How to generate a high-quality text carrying secret information is a key problem. With the widespread application of deep learning in natural language processing, recent algorithms use a language model (LM) to generate the steganographic text, which provides a higher payload compared with many previous arts. However, the security still needs to be enhanced. To tackle with this problem, we propose a novel autoregressive LS algorithm based on BERT and consistency coding, which achieves a better trade-off between embedding payload and system security. In the proposed work, based on the introduction of the masked LM, given a text, we use consistency coding to make up for the shortcomings of block coding used in the previous work so that we can encode arbitrary-size candidate token set and take advantages of the probability distribution for information hiding. The masked positions to be embedded are filled with tokens determined by an autoregressive manner to enhance the connection between contexts and therefore maintain the quality of the text. Experimental results have shown that, compared with related works, the proposed work improves the fluency of the steganographic text while guaranteeing security, and also increases the embedding payload to a certain extent.

cs.CR↗

arXiv2022

Semantic-Preserving Linguistic Steganography by Pivot Translation and Semantic-Aware Bins Coding

Linguistic steganography (LS) aims to embed secret information into a highly encoded text for covert communication. It can be roughly divided to two main categories, i.e., modification based LS (MLS) and generation based LS (GLS). Unlike MLS that hides secret data by slightly modifying a given text without impairing the meaning of the text, GLS uses a trained language model to directly generate a text carrying secret data. A common disadvantage for MLS methods is that the embedding payload is very low, whose return is well preserving the semantic quality of the text. In contrast, GLS allows the data hider to embed a high payload, which has to pay the high price of uncontrollable semantics. In this paper, we propose a novel LS method to modify a given text by pivoting it between two different languages and embed secret data by applying a GLS-like information encoding strategy. Our purpose is to alter the expression of the given text, enabling a high payload to be embedded while keeping the semantic information unchanged. Experimental results have shown that the proposed work not only achieves a high embedding payload, but also shows superior performance in maintaining the semantic consistency and resisting linguistic steganalysis.

cs.CR↗

arXiv2022

Graph Representation Learning for Spatial Image Steganalysis

In this paper, we introduce a graph representation learning architecture for spatial image steganalysis, which is motivated by the assumption that steganographic modifications unavoidably distort the statistical characteristics of the hidden graph features derived from cover images. In the detailed architecture, we translate each image to a graph, where nodes represent the patches of the image and edges indicate the local relationships between the patches. Each node is associated with a feature vector determined from the corresponding patch by a shallow convolutional neural network (CNN) structure. By feeding the graph to an attention network, the discriminative features can be learned for efficient steganalysis. Experiments indicate that the reported architecture achieves a competitive performance compared to the benchmark CNN model, which has shown the potential of graph learning for steganalysis.

cs.MM↗

arXiv2022

Exploiting Language Model for Efficient Linguistic Steganalysis

Recent advances in linguistic steganalysis have successively applied CNN, RNN, GNN and other efficient deep models for detecting secret information in generative texts. These methods tend to seek stronger feature extractors to achieve higher steganalysis effects. However, we have found through experiments that there actually exists significant difference between automatically generated stego texts and carrier texts in terms of the conditional probability distribution of individual words. Such kind of difference can be naturally captured by the language model used for generating stego texts. Through further experiments, we conclude that this ability can be transplanted to a text classifier by pre-training and fine-tuning to improve the detection performance. Motivated by this insight, we propose two methods for efficient linguistic steganalysis. One is to pre-train a language model based on RNN, and the other is to pre-train a sequence autoencoder. The results indicate that the two methods have different degrees of performance gain compared to the randomly initialized RNN, and the convergence speed is significantly accelerated. Moreover, our methods achieved the best performance compared to related works, while providing a solution for real-world scenario where there are more cover texts than stego texts.

cs.CL↗

arXiv2021

Hiding Data Hiding

Data hiding is the art of hiding secret data into a cover object such as digital image for covert communication. In this paper, we make the first step towards hiding ``data hiding'', which is totally different from many conventional works that directly embed secret data in a given cover object. In detail, we propose a novel method to disguise data hiding tools, including a data embedding tool and a data extraction tool, as a deep neural network (DNN) with an ordinary task (i.e., style transfer). After training the DNN for both style transfer and data hiding, while the DNN can transfer the style of an image to the target one, it can also hide secret data into a cover image or extract secret data from a stego image. In other words, the tools of data hiding are hidden to avoid arousing suspicion. Experimental results and analysis have shown the feasibility, applicability and superiority of the proposed method.

cs.CR↗

arXiv2021

Generating Watermarked Adversarial Texts

Adversarial example generation has been a hot spot in recent years because it can cause deep neural networks (DNNs) to misclassify the generated adversarial examples, which reveals the vulnerability of DNNs, motivating us to find good solutions to improve the robustness of DNN models. Due to the extensiveness and high liquidity of natural language over the social networks, various natural language based adversarial attack algorithms have been proposed in the literature. These algorithms generate adversarial text examples with high semantic quality. However, the generated adversarial text examples may be maliciously or illegally used. In order to tackle with this problem, we present a general framework for generating watermarked adversarial text examples. For each word in a given text, a set of candidate words are determined to ensure that all the words in the set can be used to either carry secret bits or facilitate the construction of adversarial example. By applying a word-level adversarial text generation algorithm, the watermarked adversarial text example can be finally generated. Experiments show that the adversarial text examples generated by the proposed method not only successfully fool advanced DNN models, but also carry a watermark that can effectively verify the ownership and trace the source of the adversarial examples. Moreover, the watermark can still survive after attacked with adversarial example generation algorithms, which has shown the applicability and superiority.

cs.CR↗

arXiv2021

Structural Watermarking to Deep Neural Networks via Network Channel Pruning

In order to protect the intellectual property (IP) of deep neural networks (DNNs), many existing DNN watermarking techniques either embed watermarks directly into the DNN parameters or insert backdoor watermarks by fine-tuning the DNN parameters, which, however, cannot resist against various attack methods that remove watermarks by altering DNN parameters. In this paper, we bypass such attacks by introducing a structural watermarking scheme that utilizes channel pruning to embed the watermark into the host DNN architecture instead of crafting the DNN parameters. To be specific, during watermark embedding, we prune the internal channels of the host DNN with the channel pruning rates controlled by the watermark. During watermark extraction, the watermark is retrieved by identifying the channel pruning rates from the architecture of the target DNN model. Due to the superiority of pruning mechanism, the performance of the DNN model on its original task is reserved during watermark embedding. Experimental results have shown that, the proposed work enables the embedded watermark to be reliably recovered and provides a sufficient payload, without sacrificing the usability of the DNN model. It is also demonstrated that the proposed work is robust against common transforms and attacks designed for conventional watermarking approaches.

cs.CR↗

arXiv2021

Watermarking Graph Neural Networks by Random Graphs

Many learning tasks require us to deal with graph data which contains rich relational information among elements, leading increasing graph neural network (GNN) models to be deployed in industrial products for improving the quality of service. However, they also raise challenges to model authentication. It is necessary to protect the ownership of the GNN models, which motivates us to present a watermarking method to GNN models in this paper. In the proposed method, an Erdos-Renyi (ER) random graph with random node feature vectors and labels is randomly generated as a trigger to train the GNN to be protected together with the normal samples. During model training, the secret watermark is embedded into the label predictions of the ER graph nodes. During model verification, by activating a marked GNN with the trigger ER graph, the watermark can be reconstructed from the output to verify the ownership. Since the ER graph was randomly generated, by feeding it to a non-marked GNN, the label predictions of the graph nodes are random, resulting in a low false alarm rate (of the proposed work). Experimental results have also shown that, the performance of a marked GNN on its original task will not be impaired. Moreover, it is robust against model compression and fine-tuning, which has shown the superiority and applicability.

cs.CR↗

arXiv2021

Orientation Convolutional Networks for Image Recognition

Deep Convolutional Neural Networks (DCNNs) are capable of obtaining powerful image representations, which have attracted great attentions in image recognition. However, they are limited in modeling orientation transformation by the internal mechanism. In this paper, we develop Orientation Convolution Networks (OCNs) for image recognition based on the proposed Landmark Gabor Filters (LGFs) that the robustness of the learned representation against changed of orientation can be enhanced. By modulating the convolutional filter with LGFs, OCNs can be compatible with any existing deep learning networks. LGFs act as a Gabor filter bank achieved by selecting $ p $ $ \left( \ll n\right) $ representative Gabor filters as andmarks and express the original Gabor filters as sparse linear combinations of these landmarks. Specifically, based on a matrix factorization framework, a flexible integration for the local and the global structure of original Gabor filters by sparsity and low-rank constraints is utilized. With the propogation of the low-rank structure, the corresponding sparsity for representation of original Gabor filter bank can be significantly promoted. Experimental results over several benchmarks demonstrate that our method is less sensitive to the orientation and produce higher performance both in accuracy and cost, compared with the existing state-of-art methods. Besides, our OCNs have few parameters to learn and can significantly reduce the complexity of training network.

cs.CV↗

arXiv2019

Ensemble Reversible Data Hiding

The conventional reversible data hiding (RDH) algorithms often consider the host as a whole to embed a secret payload. In order to achieve satisfactory rate-distortion performance, the secret bits are embedded into the noise-like component of the host such as prediction errors. From the rate-distortion optimization view, it may be not optimal since the data embedding units use the identical parameters. This motivates us to present a segmented data embedding strategy for efficient RDH in this paper, in which the raw host could be partitioned into multiple subhosts such that each one can freely optimize and use the data embedding parameters. Moreover, it enables us to apply different RDH algorithms within different subhosts, which is defined as ensemble. Notice that, the ensemble defined here is different from that in machine learning. Accordingly, the conventional operation corresponds to a special case of the proposed work. Since it is a general strategy, we combine some state-of-the-art algorithms to construct a new system using the proposed embedding strategy to evaluate the rate-distortion performance. Experimental results have shown that, the ensemble RDH system could outperform the original versions in most cases, which has shown the superiority and applicability.

cs.MM↗