Recently, researchers have shifted focus to reversible data hiding (RDH) schemes for JPEG images. The reinforcement learning (RL) is a solution for RDH to automatically acquire the optimal two-dimensional (2D) mapping for 2D histograms of non-zero quantized alternating current coefficients. However, merely utilizing the payload-distortion reward mechanism (PDRM) in RL cannot inject the payload guidance to the 2D mapping generation process. To tackle this issue, we propose a payload supplementary reward mechanism (PSRM) and incorporate PDRM and PSRM into RL to construct DR-2DNet, a dual-reward guided 2D mapping generation network with considering additional payload guidance. DR-2DNet generates two candidate 2D mappings, one with low distortion generated by merely utilizing PDRM and the other with low distortion and high payload obtained by jointly using PDRM and PSRM. Finally, according to the required payload, the one with the lower distortion selected from two acquired 2D mappings is used for achieving data embedding. To priorly select the frequency bands with low costs for data embedding, a frequency selection strategy combining the smoothness and embedding performance of the frequency band is designed to evaluate the cost of each frequency band, reducing image distortion and preserving the file size. Extensive experiments are conducted on the Kodak dataset and 100 images randomly chosen from the BOSSBase dataset, and the results demonstrate that the proposed method is superior to several related state-of-the-art RDH schemes for JPEG images.

Recent AI-generated image (AIGI) detectors achieve impressive accuracy under clean condition. In view of anti-forensics, it is significant to develop advanced adversarial attacks for evaluating the security of such detectors, which remains unexplored sufficiently. This letter proposes a Dual-domain Feature Importance Attack (DuFIA) scheme to invalidate AIGI detectors to some extent. Forensically important features are captured by the spatially interpolated gradient and frequency-aware perturbation. The adversarial transferability is enhanced by jointly modeling spatial and frequency-domain feature importances, which are fused to guide the optimization-based adversarial example generation. Extensive experiments across various AIGI detectors verify the cross-model transferability, transparency and robustness of DuFIA.

Copy-move forgery detection (CMFD) is a technique tailored to detect the existence of copy-move regions in a query image. In this paper, a dual-view CMFD network named DV-Net is proposed, which integrates the combination of the similarity information and tampered features from shallow features conducive to copy-move region localization by using dual-view self-correlation calculation (DV-SCC) and the shallow similarity attention module (SSAM), and strengthens the ability of distinguishing source/target regions by making deep features pass through three serial multiple serial adaptive receptive field selection modules (ARFSMs). The SCC plays an irreplaceable role in identifying copy-move regions. However, single-view SCC, such as the cosine similarity or the Euclidean distance, can solely capture the similarly information from a single perspective. DV-SCC, a combination of Euclidean distance and cosine similarity, provides more comprehensive similarity information from numerical and directional perspectives. In addition, different from previous CMFD networks that only utilize the similarity information to locate similar regions while neglecting tampered features contained in the shallow features, which are of vital importance to CMFD, we innovatively convert the similarity information into the SSAM and apply SSAM on the shallow features to emphasize the similarity information while preserving tampered features, significantly enhancing the localization accuracy of source/target regions. Multiple serial ARFSMs, each containing two parallel branches controlled by a soft attention, can adaptively select appropriate receptive fields according to the scales of tampered regions, improving the classification accuracy of source/target regions. The experimental results show that DV-Net outperforms several advanced algorithms in source/target region localization and discrimination on three publicly available datasets.

Pixel dependency modeling from tampered images is pivotal for image forgery localization. Current approaches predominantly rely on Convolutional Neural Networks (CNNs) or Transformer-based models, which often either lack sufficient receptive fields or entail significant computational overheads. Recently, State Space Models (SSMs), exemplified by Mamba, have emerged as a promising approach. They not only excel in modeling long-range interactions but also maintain a linear computational complexity. In this paper, we propose LoMa, a novel image forgery localization method that leverages the selective SSMs. Specifically, LoMa initially employs atrous selective scan to traverse the spatial domain and convert the tampered image into ordered patch sequences, and subsequently applies multi-directional state space modeling. In addition, an auxiliary convolutional branch is introduced to enhance local feature extraction. Extensive experimental results validate the superiority of LoMa over CNN-based and Transformer-based state-of-the-arts. To our best knowledge, this is the first image forgery localization model constructed based on the SSM-based model. We aim to establish a baseline and provide valuable insights for the future development of more efficient and effective SSM-based forgery localization models.

When the target images to be detected and the source images used to train the steganalyzer come from different distributions, the cover-source mismatch (CSM) occurs, which often leads to a sharp decrease in detection accuracy. To alleviate the problem, this letter proposes a four-stage steganalyzer, named ICSNet. The first three stages concentrate on generating a labeled intermediate domain to build a bridge between source/target domains. To be specific, the labeled intermediate domain is constructed by first adding the noise to target samples using a noise adding module to generate intermediate samples following the source domain distribution, and subsequently performing data embedding to these intermediate samples to generate the stego intermediate samples. The last stage focuses on strengthening the fusion of channel-wise and spatial correlations between source/target domains by presenting a coarse-to-fine two-domain channel-wise correlation fusion (CCF) module and a source-guided two-domain spatial correlation fusion (SCF) module. In CCF, the labeled intermediate domain works as a supplementary to the target domain, so that source/target domains can guide each other to reinforce the fusion of channel-wise correlations. In SCF, the labeled source domain guides the target domain to consolidate the fusion of spatial correlations. The four stages work together to reduce the distribution shift between source/target domains, thereby bringing performance improvement. Experimental results demonstrate that ICSNet significantly outperforms existing methods across various CSM scenarios.

The existing deep learning based reversible data hiding (RDH) predictors typically adopt standard convolutions for extracting features, which inherently fails to capture contextual information across different scales, making the model have difficulty to fully understand the image content. To this end, a gated multi-scale module (GMM) is proposed to enrich and strengthen feature representations by collecting multi-scale features with less computational cost using a set of parallel depthwise convolutions, and customizing the gated convolution (GConv) for RDH to weight the importance of features in channel and spatial dimensions. Considering that directly utilizing the addition or concatenation operations cannot better fuse two types of features with different receptive fields, a gated feature fusion and refinement module (GFFRM) is tailored to employ the standard convolutions of different sizes to shorten the receptive field differences between deep and shallow features. GFFRM also constructs depthwise separable convolution followed by GConv to enrich and refine the expression of features at low computational cost and enhance the information exchange across channel and spatial dimensions, thereby improving the fusion effect of features at different levels. A two-path multi-dimensional feature interaction module (MFIM) is designed, where one branch utilizes a pointwise convolution to obtain low-dimensional representations of features, whereas the other branch fuse two linearly transformed features through element- wise multiplication constructs to generate implicit high-dimensional features. GFFRM and MFIM are complementary for each other to enhance the prediction performance. Three modules, namely GMM, GFFRM and MFIM, are embedded in U -shaped encoder-decoder architecture to establish a novel RDH predictor GURNet. Extensive experiments implemented on four publicly available datasets demonstrate the superiority of GURNet, compared with state-of-the-art RDH predictors.

In general, modifying the latent diffusion model (LDM) decoder to achieve in-generation watermarking cannot introduce tremendous computational burden, which easily leads to non-convergence. This necessarily increases the difficulty of embedding the watermark into the LDM decoder due to the need to strike a balance among imperceptibility, robustness and computational cost. We realize the difficulty and design two lightweight watermarking modules, namely a low-cost watermark redundancy enhancement module (WREM) and a latent-guided watermark enhancement module (LWEM), aiming at reducing the modifications to the LDM decoder as much as possible while maintaining the generation quality and enhancing the robustness. Specifically, WREM, specially designed for shallow layers, utilizes a small number of repetition operations to strengthen the robustness of the watermark, and adopts a low-cost sub-pixel convolution layer to achieve dimension consistency between the watermark residual and the input latent, greatly reducing the computational cost while enhancing the integration of watermark features and the latent feature. LWEM, tailored for deep layers, innovatively exploits a simple bilinear interpolation to strengthen the robustness of the watermark, and fuses watermark features and the latent feature using a cheap convolution layer so as to generate the watermark residual with relatively low impact on the input latent. Combining WREM and LWEM, we construct a lightweight encoder-noiselayer-decoder in-generation watermarking method dubbed WL-WEM pursuing a satisfactory balance among three metrics including computational cost, generation quality and robustness. Experimental results also demonstrate that the proposed WL-WEM outperforms several related works in balancing three metrics.

Considering that the nature of the stego signal caused by spatial domain steganography and joint photographic experts group (JPEG) domain steganography is different, existing deep-learning steganalysis networks typically cannot work well in both spatial and JPEG domains. We propose a unified steganalysis network named ESNet to effectively preserve and identify the stego signal from spatial and JPEG domains. Specifically, dual-branch preprocessing extracts noise residuals by using fixed SRM kernels (branch 1) and randomly initialized kernels (branch 2), fuses the features from two branches and exchanges the fused complementary information through two carefully designed bidirectional fusion blocks, thereby effectively enhancing the signal-to-noise ratio. During feature extraction, considering that low-level features, such as texture and edge, are indispensable for steganalysis, we gather multi-level feature maps at different layers of the network to provide richer feature representations and merge them by using a multi-level feature fusion module, which learns the weight of different features in single-level feature map to enhance the expression of steganographic features. During classification, the multi-scale attention pooling module is employed to extract multi-scale features by designing convolution kernels of different sizes. After concatenating features of different scales, gated channel transformation is exploited to weight the importance of each channel to further strengthen the representations of steganographic features. Finally, stylepooling in combination with global standard deviation pooling and global average pooling, is used to compress channels and preserve the representation ability of channels as much as possible for classification. The experimental results show that the proposed ESNet exhibits state-of-the-art detection performance in both spatial and JPEG domains, and achieves satisfactory robustness against the cover source mismatch.

Digital video inpainting technique has been substantially improved with deep learning in recent years. It may be used as malicious manipulation to remove important objects for creating forged videos. As such it is significant to blindly identify the inpainted regions in videos. In this paper, we present a Trusted Video Inpainting Localization network (TruVIL) with excellent robustness and generalization ability. Observing that high-frequency noise can effectively unveil the inpainted regions, we design deep attentive noise learning in multiple stages to capture the inpainting traces. Firstly, a multiscale noise extraction module based on 3D High Pass (HP3D) layers is used to create the noise modality from input RGB frames. Then the correlation between such two complementary modalities are explored by a cross-modality attentive fusion module to facilitate mutual feature learning. Lastly, spatial details are selectively enhanced by an attentive noise decoding module to boost the localization performance of the network. To prepare enough training samples, we also build a frame-level video object segmentation dataset (VOS2k5) with 2500 videos and pixel-level annotation for all frames. Both quantitative and qualitative evaluations on various inpainted videos verify the robustness against video compression and generalization ability of TruVIL.

Image forgery localization, which aims to segment tampered regions in an image, is a fundamental yet challenging digital forensic task. While some deep learning-based forensic methods have achieved impressive results, they directly learn pixel-to-label mappings without fully exploiting the relationship between pixels in the feature space. To address such deficiency, we propose a Multi-view Pixel-wise Contrastive algorithm (MPC) for image forgery localization. Specifically, we first pre-train the feature extraction backbone network with a supervised contrastive loss to model pixel relationships in view of within-image, cross-scale and cross-modality. That is aimed at increasing intra-class compactness and inter-class separability. Then the localization head is fine-tuned using cross-entropy loss, resulting in a better forged pixel localizer. The MPC is trained on three different scale training datasets to make a comprehensive and fair comparison with existing image forgery localization algorithms. Extensive test results on over ten public datasets show that the proposed MPC achieves higher generalization performance and robustness than the state-of-the-arts. It is particularly noteworthy that our approach maintains a high level of localization accuracy under various post-processing combinations that approximate real-world scenarios, as well as when confronted with novel intelligent editing techniques. Finally, comprehensive and detailed ablation experiments demonstrate the reasonableness of MPC.

The commonly used standard convolutional layers cannot adaptively adjust the number and locations of sampling points according to the scales and shapes of tampered regions, which increases the difficulty of detecting images containing tampered regions of different sizes. Therefore, the selective sampling attention (SSA) is proposed to automatically learn the number and locations of sampling points as well as the weight of each sampling point within a certain context range of the input feature map through backpropagation, which can help the network better adapt to tampered regions of different scales and shapes. In addition, the self-correlation calculation (SCC), aiming at calculating the similarity between every two feature points in a feature map, necessarily incurs an expensive computational burden when used for high-resolution feature maps. To remedy the problem, the two-step SCC (TS-SCC) with low computation burden is proposed to pick out highly similar regions by means of the feature similarity obtained from low-resolution version of the input feature map, so that the high-resolution version merely needs to calculate the similarity between every two feature points within its high-similarity regions. Finally, to predict the edges and interiors of copy-move tampered regions more precisely, adaptive dual-branch feature fusion module is proposed to employ a lightweight multi-scale atrous convolutional module to adaptively fuse multi-level features before TS-SCC and the correlation features after TS-SCC, thereby improving the detection performance. Combining these three structures, a lightweight, fast, low-cost and high-precision CMFD network, ST-Net, is designed in this paper. Experimental results on four publicly available datasets verify that ST-Net outperforms several related CMFD networks in terms of detection accuracy, number of parameters, computational cost and inference time.

Essentially, directly introducing any object detection network to perform copy-move forgery detection (CMFD) inevitably leads to low detection accuracy. Therefore, DCM-Net, an object detection network dominated by diffusion model that incorporates the characteristics of copy-move forgery, is proposed in this paper for obviously enhancing CMFD performance. DCM-Net, as the first diffusion model-based CMFD network, has the following three improvements. Firstly, the high-similarity box padding strategy pads high-similarity boxes, rather than random boxes used in diffusion model, to ground truth boxes, better guiding subsequent dual-attention detection heads (DDHs) to focus more on high-similarity regions. Secondly, different from previous deep learning based CMFD networks that utilize self-correlation calculation to indiscriminately transform all classification features extracted from feature extraction into high-similarly features, an adaptive feature combination strategy is proposed to obtain the optimal feature transformation capable of achieving the best detection performance, enabling DDHs to more effectively distinguish source and target regions. Finally, to make detection heads have more accurate source/target localization and distinguishment, DDHs equipped with efficient multi-scale attention and contextual transformer, are proposed to generate tampered features fusing the entire precise spatial position information and rich contextual global information. The experimental results carried out on three publicly available datasets including USC-ISI, CoMoFoD, and COVERAGE, demonstrate that DCM-Net outperforms several advanced algorithms in terms of similarity detection ability and source/target differentiation ability.

For the prediction unit partition modes (PUPM)-based steganography, a mainstream branch of high efficiency video coding (HEVC) video steganography, striking a balance between embedding performance and security is very challenging. Including the$2\mathcal {N} \times 2\mathcal {N}$PUPMs having the maximum number of PUPMs into data embedding is indeed an effective way of enlarging the embedding capacity, but it necessarily causes a significant decline in security. Therefore, a multi-factor-involved cost function (MFICF) is proposed in this paper to evaluate the embedding cost for modifying each PUPM by comprehensively considering four different aspects affecting the embedding performance and security. With the assistance of MFICF, the 7-ary notational system is combined to use all the 7 types of PUPMs containing$2\mathcal {N} \times 2\mathcal {N}$for data embedding, thus enlarging the embedding capacity as well as enhancing the embedding efficiency. The syndrome-trellis code driven by MFICF, named CFSTC, is designed to preferentially select PUPMs with low embedding costs for data embedding, so that the embedding efficiency is largely enhanced. The security is effectively guaranteed by allocating a large embedding cost for modifying$2\mathcal {N} \times 2\mathcal {N}$to another type of PUPM. Finally, a lightweight convolutional neural network in combination with gated channel transformation, called GSCNet, is proposed to replace the in-loop filter in HEVC, further optimizing the visual distortion and bitrate increase caused by data embedding. Combining these components above, we design a PUPM-based steganography algorithm, GSAPM. Experimental results show that GSAPM effectively enhances the embedding performance while maintaining high security.

This letter proposes GDNet equipped with the generation of discriminative mixing regions (GDMR) and discriminability-aware local image mixing (DLIM), a steganalysis network aiming at alleviating significant accuracy degradation caused by cover-source mismatch (CSM), which pertains to the situation where source and target domains come from different distributions. GDNet guides a steganalyzer trained on the source domain to the target domain by mixing the source and target images at the region-level and pixel-level to construct a discriminative intermediate domain. On the one hand, GDMR designs an epoch-related region-level mixing ratio to control the size of the mixed region, and based on this ratio, selects the regions within the target image strongly related to the stego signal to participate in the generation of the intermediate domain, while suppressing other regions weakly related to the stego signal. On the other hand, DLIM utilizes the pixel-level mixing ratio to reduce the impact of the regions weakly related to the stego signal on the discriminability of the intermediate domain as the region-level mixing ratio increases, thereby increasing the diversity of the intermediate domain. Experimental results demonstrate that GDNet significantly outperforms existing methods across various CSM scenarios.

Existing convolutional neural network-based reversible data hiding (RDH) predictors typically stack the standard convolution blocks with stride 1 for feature extraction, and keep the sizes of input and output feature maps unchanged through padding. This suggests that only a limited range of contextual spatial information is obtained. To remedy this problem above, a U-Net-like RDH predictor named UCANet is proposed in this paper to capture rich multi-scale contextual information by gradually downsampling feature maps. To fuse two feature maps at different levels along the channel dimension, we put forward the channel adaptive attention (CAA). By merely combining cheap pointwise convolution operations, CAA achieves the integration of non-linear and linear features as well as implicitly enhances channel dimensionality with low computational burden, thereby effectively enriching the expression of the channel information. The design of UCANet considers the characteristics of RDH from two aspects. On the one hand, instead of maxpooling or average pooling commonly used for downsampling, a stride-2 convolution block that can adaptively adjust the weights of convolution kernels and select useful information is utilized to downsample feature maps. On the other hand, UCANet removes the batch normalization layers to avoid their influence on the distribution of feature maps, which helps to strengthen the network's prediction capability. Extensive experiments also demonstrate that the proposed UCANet achieves better prediction performance, compared to several state-of-the-art methods.

The existing deep learning based copy-move forgery detection (DL-CMFD) networks focus on providing impressive detection accuracy for tampered regions of different sizes, but usually result in high computation cost and a large number of parameters. The focus of this letter is to propose LHCM-Net, a lightweight, high-precision DL-CMFD network, by integrating a low-cost self-correlation calculation (SCC) module (LCSCC), a gated feature fusion module (GFFM) and residual U-blocks (RSU) equipped with FasterNet blocks (FRSU). Considering that SCC, which calculates the similarity between every two pixels, inevitably leads to high computation cost, existing DL-CMFD networks have to carry out SCC only on low-resolution feature maps to reduce the computation cost. To make high-resolution features available for SCC without obviously introducing high computation cost, this letter proposes LCSCC to calculate the similarity between pixels with a certain distance. GFFM is presented to fuse feature maps of different spatial resolutions by adaptively adjusting their weights based on their respective characteristics, thereby fully integrating high-resolution and low-resolution features for subsequent LCSCC and obviously enhancing the detection accuracy. The FRSU allows LHCM-Net to keep the number of parameters (NP) and computation cost low by combining lightweight FasterNet blocks. The experimental results also demonstrate that LHCM-Net outperforms several existing DL-CMFD networks on three publicly available datasets in terms of detection accuracy, NP and computation cost.

In general, cover source mismatch (CSM) inevitably leads to a significant decrease in detection accuracy in image steganalysis because the source and target domains have different distributions. To remedy this problem, a universal mismatched steganalyzer ISNet equipped with generated local mixing positions, a local feature-level mixup-related patchup (LFMP), and domain factors is proposed for both spatial and JPEG domains in this paper. Unlike existing deep steganalysis networks, which simply minimize the domain discrepancy between source and target to address the problem of CSM, and thus, cannot handle large domain discrepancy, IDGM based on LFMP generates diverse intermediate domains to bridge the two extreme domains so as to alleviate the decrease in detection accuracy caused by CSM. Moreover, ISNet enables the intermediate domain distribution to progressively transit from source to target by adjusting the domain factor sampled from Beta(α, 1), so that the classifier gradually adapted to target provides an improvement in discriminability on the target domain. The experimental results show that ISNet achieves the best performance in various CSM cases, compared with the most advanced deep learning-based steganalysis network.

Night-time image enhancement (NIE) aims at boosting the intensity of low-light regions while suppressing noises or light effects in night-time images, and numerous efforts have been made for this task. However, few explorations focus on the quality evaluation issue of enhanced night-time images (ENTIs), and how to fairly compare the performance of different NIE algorithms remains a challenging problem. In this paper, we firstly construct a new Real-world Night-Time Image Enhancement Quality Assessment (i.e., RNTIEQA) dataset that includes two typical types of night-time scenes (i.e., extremely low light and uneven light scenes), and carry out human subjective studies to compare the quality of ENTIs obtained by a set of representative NIE algorithms. Afterwards, a new objective ranking method that comprehensively considering image intrinsic and impairment attributes is proposed for automatically predicting the quality of ENTIs. Experimental results on our RNTIEQA dataset demonstrate that the proposed method outperforms the off-the-shelf competitors. Our dataset and code will be released athttps://github.com/Leilei-Huang-work/RNTIEQA-dataset.

Copy-move forgery causes a big challenge to copy-move forgery detection (CMFD) due to that the photometrical characteristics of genuine and tampered regions in the same image remain highly consistent. A novel U-Net-like architecture with multiple asymmetric cross-layer connections associated with self-correlation and atrous spatial pyramid pooling (ASPP) between feature extraction module (FEM) and tampered region localization module (TRLM), called UCM-Net, is proposed in this article. Different from existing deep learning based CMFD networks which indiscriminately process large or small tampered regions without considering the statistical characteristics of regions, FEM differentially treats large or small tampered regions by exploiting deep backbone networks to extract high-level features with rich semantic information for large tampered regions while utilizing lightweight backbone networks to extract low-level features for small tampered regions. Multiple cross-layer connections between two modules utilize the self-correlation calculation and ASPP to remove as much irrelevant semantic information as possible while retaining multi-scale tampered features from shallow to deep convolutional layers of FEM. Unlike the previous CMFD networks, which cannot capture multi-scale features because of simply stacking convolution blocks in the upsampling step, TRLM exploits multiple U-shaped residual U-block modules with different depths to change the receptive field of each point in the tampered feature maps so as to capture global and local information, greatly improving the localization accuracy of tampered regions. Experimental results on three publicly available databases demonstrate that UCM-Net outperforms several state-of-the-art algorithms in terms of various evaluation metrics.

Recently, reversible data hiding in encrypted images (RDHEI) has received widespread attention from researchers. To embed high payload into encrypted images while maintaining sufficient security, a novel RDHEI algorithm in combination with consecutive zero-valued high bit-planes compression, bit-plane swapping as well as block rearrangement is proposed in this article. The proposed method is the first work to compress global zero-valued high bit-planes in a block-wise manner and adaptively allocate different Huffman indicators based on the occurrence frequency of zero-valued bit-planes so that a higher embedded payload is greatly provided. Unlike existing RDHEI methods embedded with unencrypted auxiliary information, resulting in low security, the bit-plane swapping and block rearrangement are subtly designed to cluster together all embeddable bit-planes, which enables most auxiliary information to be encrypted, largely enhancing the security and facilitating data embedding and data extraction. The experiment results demonstrate that the proposed method outperforms some state-of-the-art RDHEI methods in terms of security and payload. The average payload of the proposed method for two publicly-used datasets including BOSSbase and BOWS-2, are 3.793 bpp and 3.705 bpp, respectively.

Effectively enhancing the weak stego signal while striking the balance among three evaluation metrics, i.e., detection accuracy, time cost, as well as the number of parameters (NP) is indeed a huge challenge for existing deep learning-based steganalysis detectors. In this letter, a novel steganalysis detector called GFS-Net is proposed, aiming at enhancing the stego signal as much as possible while balancing the three metrics to obtain comprehensive detection performance. In preprocessing, combining highly lightweight gated channel transformation with a pointwise convolution layer used for enlarging the number of channels enriches the expression of the stego signal by promoting cooperation among enlarged channels, thereby significantly improving the signal-to-noise ratio while avoiding occupying a large NP. Moreover, two FasterNet blocks equipped with partial convolution having a small NP, rather than residual blocks, are applied to the last two layers of feature extraction to efficiently extract the stego signal by reducing the calculation of similar features, so that the computational cost and NP are saved. Finally, compared to using the global average pooling (GAP) alone, the stylepooling jointly utilizing the global standard deviation pooling and GAP helps the subsequent fully connected better identify weak stego signal, and thus improves detection accuracy. By means of the above three perspectives, GFS-Net with only 0.13 M parameters is obtained. Experimental results also demonstrate that GFS-Net achieves higher detection accuracy and lower computational cost than state-of-the-art steganalysis detectors.

This letter aims to propose an advanced predictor called NU$^{2}$P in the multiple histogram modification (MHM)-based framework by combining deep learning techniques that take the local characteristics of pixels into account. NU$^{2}$P is firstly designed for reversible data hiding by incorporating U$^{2}$P (an improved version of U$^{2}$-Net) equipped with convolutional block attention module (CBAM). The purpose of U$^{2}$P is to integrate the feature maps with different sizes and receptive fields through U-Net-like residual U-blocks (RSUs), which can make full of strong correlations between adjacent pixels while reducing the computational cost by means of the pooling layers of RSUs. CBAM pays different attention to feature maps and elements of feature maps from the perspectives of channel and spatial attention, thereby helping NU$^{2}$P further enhance the prediction performance. In the MHM-based framework, for multiple categories generated using fuzzy C-means with multiple deliberately-designed features, NU$^{2}$P is conductive to constructing a sharp prediction error histogram (PEH) for each category and the improved discrete particle swarm optimization without significantly increasing the computational cost is used to adaptively select the optimal bins for each PEH. The experimental results show that the proposed method significantly outperforms several state-of-the-art RDH methods in terms of image quality and payload.

In this letter, a novel deep steganalysis network called SwT-SN is proposed by integrating directional difference adaptive combination (DDAC) followed by three residual blocks, convolutional spatial pyramid pooling equipped with size-independent detector (CSPP-SID) as well as a two-part of Swin transformer (SwT) structure suitable for steganalysis, aiming at enhancing the detection accuracy for arbitrary-sized images while significantly reducing training and test cost. In comparison to simply exploiting DDAC as preprocessing, DDAC + the residual structure can better improve the signal-to-noise ratio of the residual maps by suppressing the image content. In addition, CSPP-SID is innovatively proposed to convert feature maps of any size into feature vectors with fixed dimension, helping SwT-SN achieving high detection accuracy for arbitrary-sized images. Finally, a two-part structure of SwT with a fixed number of patches is firstly designed for image steganalysis to greatly reduce training cost by calculating the multi-head self-attention mechanism in the shifted window. Extensive experiments on two benckmark databases verify that SwT-SN has higher detection accuracy and shorter training cost compared to two prior state-of-the-art networks.

In this paper, a general reversible data hiding (RDH) framework for joint photographic experts group (JPEG) images with multiple two dimensional histograms (2DHs) is proposed. Regardless of whether zero alternating current (AC) coefficients are included to join data embedding or only non-zero AC coefficients are applied, the performance in terms of visual quality and file size increment is improved by using the proposed framework. This framework is mainly composed of the following three parts: histogram generation, adaptive 2DH mapping selection, and improved discrete particle swarm optimization (IDPSO). Unlike existing 2DH-based JPEG RDH methods, in which a uniform threshold is utilized to construct multiple histograms, in histogram generation, thresholds for different histograms are adaptively assigned according to the local properties of histogram coefficients. As a result, as many coefficients in complex regions as possible are excluded from the construction of each histogram. We subtly design multiple 2DH mappings, and adaptively select 2DH mappings for different 2DHs based on their distribution characteristics. Through slight adjustments, each 2DH mapping can be employed in cases where either zero AC coefficients or only non-zero AC coefficients are used for data embedding. Adaptive threshold and 2DH mapping selection provide a better image quality at a given embedding capacity but inevitably cause considerable complexity cost. To significantly reduce the computational cost, we propose IDPSO by combining differential evolution. IDPSO has the advantages of rapid convergence speed as well as satisfactory qualities of the best solutions. With the help of differential evolution, IDPSO expands the diversity of particles and efficiently avoids local optimal trapping problems. The experimental results also demonstrate the effectiveness of the proposed method in terms of visual quality, file size increment and complexity cost.

Reversible data hiding based on joint photographic experts group (JPEG) images has been extensively studied to enhance embedding performance in terms of visual quality and file size preservation at the desired payload. In this paper, an efficient adaptive RDH method for JPEG images with multiple two-dimensional (2D) histogram modification is proposed. Firstly, the proposed method proposes the block smoothness estimator and the band smoothness estimator, and then combines the two estimators to reduce the embedding distortion as much as possible at the desired payload. Instead of adopting a fixed 2D mapping or choosing one from several empirically-designed mappings for each 2D histogram, the proposed method designs an adaptive 2D mapping generation strategy to adaptively generate a large number of mappings with considering the local characteristics of histogram distribution. Since exhaustively searching for the optimal mapping achieving the highest embedding performance for each 2D histogram is time-consuming, an improved discrete particle swarm optimization is utilized in the proposed method to speed up the optimization process. Extensive experimental results also demonstrate the effectiveness of the proposed method in terms of visual quality and file size increment of the stego image.

In this letter, a lightweight and effective deep steganalysis network (DSN) with less than 400,000 parameters, called LWENet, is proposed, which focuses on increasing the performance as well as significantly reducing the number of parameters (NP) from three perspectives. Firstly, in the preprocessing part, several lightweight bottleneck residual blocks are combined into the spatial rich model filters to improve the signal-to-noise ratio of stego signals while slightly increasing NP, thereby improving the subsequent performance. Secondly, a depthwise separable convolution layer is exploited at the end of the feature extraction part to largely reduce NP and increase the performance by capturing salient correlations while ignoring trivial ones among feature maps. Finally, to keep LWENet lightweight, we have to select only one fully connected (FC) layer. Simultaneously, multi-view global pooling is employed prior to the FC layer to yield multi-view features and further improve the detection performance. Extensive experiments demonstrate that our network achieves better performance than several state-of-the-art DSNs.

Reversible data hiding with contrast enhancement (RDH-CE) is proposed to aim at improving the contrast of images while embedding data. After deeply analyzing and studying the RDH-CE method proposed by Jafaret al., it is found that there are three main problems in their method. Firstly, their method ignores the fact that the left-bottom neighbors of a pixel contribute to increasing the accuracy of the local-complexity evaluation. Secondly, Jafaret al.’s method employs K-means clustering in combination with one single feature to split pixels into five classes, leading to a weak clustering performance. Finally, Jafaret al.’s method uniformly embedded 1 bit into each pixel irrespective of the local complexity, and thus, the embedding capacity is limited. To this end, an improved RDH-CE method is proposed in this paper. Considering that the complexity evaluation plays a vital role in both contrast enhancement and payload increase, we improve embedding performance by including left-bottom neighbors of a pixel into complexity evaluation. Compared with one single feature in Jafaret al.’s method, we extract multiple features to assist K-means clustering such that a better cluster performance is obtained. In addition, our method provides an adaptive pixel modification strategy based on the local complexity, in which we can adaptively embed 1 or 2 bits into a pixel according to the corresponding complexity. By these three improvements, our method is capable of achieving high capacity while enhancing contrast. The experimental results also show that our method achieves higher accuracy of the complexity evaluation, larger payload, and better local contrast enhancement than those existing RDH-CE related methods.

A reversible watermarking algorithm based on piecewise modification operation(PMO) of triplets is proposed in this paper. PMO is designed to embed a bit into any two pixels while remaining the intensity sum of pixels unchanged. Two bits are embedded into a triplet by repeatedly using PMO on every two neighboring pixels. Another advantage of using PMO is that the capacity consumed by the additional information can be largely decreased. As a result, the embedding capacity is considerably increased. A series of experiments is conducted to verify effectiveness and advantages of the proposed approach.

A novel reversible data hiding scheme based on invariability of the sum of pixel pairs and pairwise difference adjustment (PDA) is presented in this letter. For each pixel pair, if a certain value is added to one pixel while the same value is subtracted from the other, then the sum of these two pixels will remain unchanged. How to properly select this value is the key issue for the balance between reversibility and distortion. In this letter, half the difference of a pixel pair plus 1-bit watermark has been elaborately selected to satisfy this purpose. In addition, PDA is proposed to significantly reduce the capacity consumed by overhead information. A series of experiments is conducted to verify the effectiveness and advantages of the proposed approach.

A novel reversible data hiding scheme based on an integer transform is presented in this paper. The invertible integer transform exploits the correlations among four pixels in a quad. Data embedding is carried out by expanding the differences between one pixel and each of its three neighboring pixels. However, the high hiding capacity can not be achieved only by difference expansion, so the companding technique is introduced into the embedding process so as to further increase hiding capacity. A series of experiments are conducted to verify the feasibility and effectiveness of the proposed approach.

A novel reversible data hiding scheme is proposed in this article. Each pixel is predicted by its right neighboring pixel in scan order to get its prediction-error. Then, a companding technique is introduced so as to largely increase the number of prediction-errors available for embedding. Accordingly, a location map recording available positions can be compressed into a short bitstream. By largely decreasing the capacity consumed by the compressed location map, the high hiding capacity is achieved. A series of experiments are conducted to verify the feasibility and effectiveness of the proposed approach.

A reversible digital watermarking scheme based on a blockwise difference expansion (DE) method is proposed here. The proposed scheme differs from previous ones in its robustness against cropping and collage attacks. In order to easily detect the correct cropped position for a cropped image, a locating pattern is embedded into the carrier image by a modified patchwork algorithm. Incurred overhead, payload, and authentication information are then collected and embedded into the carrier image by a blockwise DE method. Within the proposed scheme, image ID and block index also serve as parts of the inputs to the hash function. This endows the scheme with capability in resisting collage attacks. The experimental results show that the scheme can achieve a good robustness against cropping attacks as well as collage attacks.