Wei Wu

With the widespread deployment of the Internet of Things (IoT) and distributed sensing systems, visual data are frequently transmitted over bandwidth-limited and unreliable links, where compression and noise degradation are inevitable. This poses significant challenges to secure, imperceptible, and reversible image communication. Existing steganographic methods often suffer from limited reversibility, noticeable statistical distortion, and weak resistance to steganalysis under realistic channel conditions. This paper proposes Rivic, a reversible image hiding framework for robust and secure visual communication. Rivic integrates wavelet-domain multi-frequency modeling, hierarchical spectral refinement, multi-attentive correction, and multi-domain synergistic recovery into an end-to-end invertible network, aiming to enhance embedding stability and noise-resilient reconstruction. Experimental results on ImageNet, COCO, and ImageHide demonstrate that Rivic achieves 34–35 dB PSNR and 0.93–0.95 SSIM in the cover–stego task, yielding a 5–7 dB PSNR improvement over the strongest CNN-based baselines. Under various channel degradations, including additive noise and JPEG compression, Rivic consistently maintains reconstruction quality above 30 dB. Moreover, when evaluated against advanced steganalysis networks such as SRNet, Yedroudj-Net, and SiaSteg-Net, Rivic reduces the average detection accuracy to 0.50–0.62, approaching the level of random guessing. These results indicate that Rivic enables high-fidelity, robust, and secure reversible visual communication in IoT sensing and control systems operating under practical transmission constraints.

With the advent of intelligent technologies, miscellaneous data containing sensitive information are explosively generated and shared. Compressive sensing methods are naturally suitable for such scenarios due to their joint compression and encryption capabilities. However, data users of most existing compressive sensing methods need to reconstruct original images before use, which brings two disadvantages. First, indiscriminately requiring every data user to reconstruct images without considering their exact requirements is neither advisable nor efficient. Second, allowing all data users to reconstruct original images may cause private or confidential information exposure. To address these issues, in this paper, a novel image sharing method is proposed, which realizes efficient multilevel privacy preservation. Specifically, data owners compress the original images with designed measurement matrices through the proposed Tℓ₁-B2DLDA algorithm, which outputs dimension-reduced data with the ability to simultaneously support the subsequent classification tasks for level I data users and reconstruction tasks for level II data users. Therefore, low level data users could achieve their goals without obtaining any private or confidential information in the original images. Experiments are conducted to verify the feasibility, performance and robustness of the proposed method. Furthermore, the security of the proposed method is analyzed both theoretically and practically. The source code of the proposed method is publicly available at https://github.com/xchuxiao23/TL1-B2DLDA.

Privacy-preserving truth discovery (PPTD) has garnered significant attention in mobile crowdsensing (MCS). However, existing research lacks sufficient privacy protection and is often vulnerable to collusion attacks among malicious participants. Moreover, incorrect data submitted by unreliable users and their weights may reduce the accuracy of truth discovery. To address these issues, this paper proposes a privacy-preserving truth discovery framework resistant to collusion and reputation fraud (TD-RCRF) that is highly resistant to collusion and reputation fraud. The scheme employs additive secret sharing to protect sensing data, weights, intermediate results, and ground truth. To screen trustworthy users who meet reputation requirements under the non-colluding dual-server model, we propose a privacy-preserving reputation verification algorithm that combines Pedersen commitment and zero-knowledge proof to verify the validity of mobile users’ reputation values. Additionally, we propose a homomorphic strategy that converts shares between multiplication and addition and use it to design a lightweight truth discovery algorithm that further improves the accuracy of the “truth” using reputation values. Security analysis proves that TD-RCRF is privacy-preserving and secure under the non-colluding dual-server assumption. Theoretical analysis and experiments show that it is practical and efficient.

Deep-learning-based compressed sensing (CS) techniques have exhibited exceptional prowess in signal reconstruction and data-sharing applications, particularly within the realm of Internet of Things sensor data processing. However, existing methods overlook a critical security vulnerability: the susceptibility of CS techniques to backdoor attacks during the reconstruction phase, which could pose severe security risks to downstream applications. This study pioneers an investigation into the feasibility of backdoor injection during the reconstruction phase, presenting the stealthy multiconstraint adversarial backdoor network against compressed learning (MACS-BNet) and substantiating its efficacy in subverting downstream classification tasks. MACS-BNet synergistically incorporates detailed sensing enhancement, fortified by local information relative positional encoding (LiRPE), to elevate image reconstruction fidelity. Concurrently, it employs a multiconstrained adversarial optimization that integrates sparsity, amplitude regulation, and spatial smoothness constraints, achieving an optimal tradeoff between perturbation imperceptibility and attack efficacy. Consequently, victim models are subtly manipulated to yield outputs consistent with the attacker’s objectives. Extensive empirical evaluations reveal that MACS-BNet consistently surpasses seven cutting-edge attack methodologies across attack success rate (ASR), clean sample classification accuracy, and stealthiness under both all-to-one and all-to-all attack paradigms. Specifically, MACS-BNet attains an unparalleled clean classification accuracy of 99.52% and an ASR of 99.43% in the all-to-one mode, while simultaneously ensuring high-quality image reconstruction. Furthermore, MACS-BNet exhibits formidable resistance against detection by seven state-of-the-art defense mechanisms, underscoring its superior stealth and robustness.

With the rapid development of the Internet of Things (IoT), a substantial number of mobile devices may need to be connected to and communicate with the other devices. IoT smart nodes, which play a key role in these processes, provide not only connectivity but also data security during data collection and transmission. While existing research has considered security issues concerning data collection and transmission to some extent, the inherent vulnerabilities of smart nodes themselves have rarely received adequate attention. Actually, smart nodes can be susceptible to malware attacks, and the fundamental task to protect them is the timely and accurate identification of malware. However, the stealthy nature of malware, the imbalance in its types, and the sharp increase in malware variants make this task challenging. Existing static malware classification methods, which require known malware features, may struggle with a whole new set of malware variants. Dynamic methods plausibly have the potential to identify malware variants but necessitate running malware beforehand. In this article, we propose a deep-learning-based malware classification method designed to help most smart IoT nodes running mainstream operating systems preventing attacks from the malware and its variants. The proposed method incorporates an automatically adjusting threshold mechanism to extract trivial features of malware, refining the classification process even if original features are modified. It also performs well with insufficient malware samples and is compatible with most mainstream convolutional neural network (CNN) models allowing flexible implements. Experiments concerning byte-level variants, operation-code-level variants, and noise-caused variants are conducted. The results show that for the malware files and their variants, the proposed method can accurately accomplish the malware classification tasks. In addition, the proposed method offers a robust solution for the malware classification, which helps to enhance the security of smart nodes against a wide range of malware and its variants. Last but not least, its compatibility with mainstream CNN models and flexibility in deployment make it a practical approach for real-world applications.

With the rapid development of the Internet of Things (IoT), the sophistication and intelligence of sensors are continually evolving, playing increasingly important roles in smart homes, industrial automation, and remote healthcare. However, these intelligent sensors face many security threats, particularly from malware attacks. Identifying and classifying malware is crucial for preventing such attacks. As the number of sensors and their applications grow, malware targeting sensors proliferates. Processing massive malware samples is challenging due to limited bandwidth and resources in IoT environments. Therefore, compressing malware samples before transmission and classification can improve efficiency. Additionally, sharing malware samples between classification participants poses security risks, necessitating methods that prevent sample exploitation. Moreover, the complex network environments also necessitate robust classification methods. To address these challenges, this paper proposes CSMC (Compressed Sensing Malware Classification), an efficient malware classification method based on compressed sensing. This method compresses malware samples before sharing and classification, thus facilitating more effective sharing and processing. By introducing deep learning, the method can extract malware family features during compression, which classical methods cannot achieve. Furthermore, the irreversibility of the method enhances security by preventing classification participants from exploiting malware samples. Experimental results demonstrate that for malware targeting Windows and Android operating systems, CSMC outperforms many existing methods based on compressed sensing and machine or deep learning. Additionally, experiments on sample reconstruction and noise demonstrate CSMC’s capabilities in terms of security and robustness.