Crowdsensing has gained widespread attention due to its efficient and low cost data collection mode that leverages a large number of intelligent mobile devices. Privacy and data quality are two key concerns in crowdsensing. Recently, extensive efforts have been devoted to privacy-preserving truth discovery (PPTD), which aims to protect sensitive data while improving data quality. However, existing PPTD schemes suffer from either low reliability—especially under collusion attacks and server dropout—or low communication efficiency. As a result, they fail to meet the practical requirements of real-time processing with high reliability. To address the problems, we propose FPTD, a super fast PPTD scheme for crowdsensing that provides$T$-out-of-$N$threshold reliability, with a focus on boosting online efficiency. Our scheme employs edge nodes as servers in a multi-server architecture, resisting up to ($T-1$) colluding servers and ($N-T$) server dropouts. We construct novel protocols for PPTD, including secure division and negative of approximate logarithms. We further significantly improve communication efficiency by designing protocolsmultiply-then-divide,dot-product-then-divide, andfilter-then-dot-product-then-divide, all of which require only a single element per party in online communication. Our tradeoff is a need for a circuit-dependent offline phase, which is independent of the parties' inputs. Compared to the state-of-the-art scheme, we are 13$\sim 217\times$faster (LAN) and 17$\sim 190 \times$faster (WAN) in online execution time.

The right to be forgotten, mandated by modern privacy regulations, poses unique challenges to federated learning in medical imaging, where models must erase patient-specific information while preserving diagnostic utility. Existing federated unlearning (FU) methods struggle with this trade-off due to the high privacy sensitivity and spatial complexity of medical data. To address these challenges, we propose LAFUL, a lesion-aware federated unlearning framework designed for privacy-preserving medical image analysis. LAFUL exploits the spatial sparsity and semantic focus of pathological lesions to identify and selectively mask lesion-sensitive gradient channels, effectively removing private information while retaining anatomy-relevant representations. To mitigate distributional shifts caused by targeted gradient removal, a feature-level knowledge distillation module aligns intermediate representations using an unlabeled proxy dataset. Experiments on two public benchmarks—intracranial hemorrhage detection and skin lesion classification—show that LAFUL achieves near-retrained accuracy with over$9 \times$efficiency improvement, providing an interpretable and practical solution for scalable FU in healthcare.

YOLO-based models are widely used for personal protective equipment (PPE) compliance detection due to their excellent detection performance and efficiency. However, most YOLO models are not competent for detection tasks in complex industrial scenarios such as remote surveillance and extremely small targets. In addition, there is a lack of effective model lightweighting and knowledge transfer approaches for industrial deployment. To this end, this paper proposes a Multi-scale and Knowledge-Distilling YOLO (MKD-YOLO) based on YOLOv8n for efficient PPE compliance detection. Specifically, in backbone stage, we design an Efficient Multi-Scale Enhanced Convolution (C2f-EMSEC) module and Large Spatial Pyramid Pooling-Fast (LSPPF) module for multi-scale and global-contextual feature learning as well as reducing model complexity. Then, in neck stage, a refined Bidirectional feature Pyramid Network (BPNet) is designated to capture fine-grained details for extremely small object detection. Moreover, we apply channel-wise knowledge distillation to facilitate model lightweighting and domain-specific knowledge transfer learning. Experiments on our proposed dataset and public datasets show that the proposed MKD-YOLO achieves a new state-of-the-art (SOTA) detection performance and efficiency for practical PPE compliance detection tasks. Codes and the dataset are available at https://github.com/z1Zjt/MKD-YOLO.

The convergence of healthcare and financial technology has driven the adoption of federated learning (FL) for collaborative analysis of sensitive data across distributed systems. However, existing approaches face critical challenges, particularly gradient inversion attacks that can reconstruct raw patient data from shared parameters, compromising clinical confidentiality. While FL integrated with conditional generative adversarial networks (GANs) shows promise for medical applications, sharing generator parameters still poses substantial privacy risks. As AI technologies proliferate and the demand for telemedicine expands, the need for secure medical fintech solutions increases, requiring robust mechanisms to protect shared parameters and medical data. To address these limitations, we propose CryptoGAN, a novel approach that embeds a GAN into the client’s local network, aligns the generator’s output distribution with the feature distribution of local data, and aggregates the client’s information by uploading homomorphically encrypted generator parameters. This approach ensures that not only can leakage of local data features be prevented but also the sensitive medical information embedded in the generator parameters is protected, thus improving the privacy of medical applications. Extensive experiments on medical datasets demonstrated that CryptoGAN effectively protects patient privacy while maintaining high diagnostic accuracy and outperforms traditional FL methods in the healthcare domain.

Truth discovery is an effective technique for resolving data conflicts in crowdsensing. Fog-based mobile crowdsensing utilizes low-latency and high-efficiency communications capabilities of fog computing to achieve large-scale data sensing at a low cost. Privacy-preserving truth discovery (PPTD) has garnered significant attention in recent years due to the inclusion of users’ sensitive information in sensory data. However, existing PPTDs have not adequately addressed fog servers’ reliability and mobile devices’ efficiency simultaneously. Challenges are that fog servers are susceptible to breakdowns and collusion that causes privacy breaches, while mobile devices have limited resources. We thus propose a reliable and efficient PPTD for fog-based crowdsensing. We employ a threshold secret sharing scheme to establish secure multi-party computation primitives. These primitives are then used to construct an arithmetic circuit–an essential component of the PPTD. This approach preserves privacy of sensory data, as well as intermediate and final results, while accounting for server collusion, dropout, and mobile devices’ efficiency. It has$T$-out-of-$N$threshold reliability that resists ($T-1$) servers’ collusion and ($N-T$) servers’ dropout. Experimental results demonstrate that our scheme reduces worker processing time by at least one order of magnitude and network communication overhead by approximately two orders of magnitude compared to existing PPTD methods.

Single domain generalization (SDG) aims to learn a generalizable model from only one source domain available to unseen target domains. Existing SDG techniques rely on data or feature augmentation to generate distributions that complement the source domain. However, these approaches fail to address the challenge where gradient conflicts from synthesized domains impede the learning of domain-invariant representation. Inspired by the concept of mechanical equilibrium in physics, we propose a novel conflict-aware approach named domain gradient equilibrium for SDG. Unlike prior conflict-aware SDG methods that alleviate the gradient conflicts by setting them to zero or random values, the proposed domain gradient equilibrium method first decouples gradients into domaininvariant and domain-specific components. The domain-specific gradients are then adjusted and reweighted to achieve equilibrium, steering the model optimization toward a domain-invariant direction to enhance generalization capability. We conduct comprehensive experiments on four image recognition benchmarks, and our method achieves an accuracy improvement of 2.94% in the PACS dataset over existing state-of-the-art approaches, demonstrating the effectiveness of our proposed approach.

Human intelligence tasks (HITs) are widely utilized for crowdsourcing human knowledge, such as labeling images for machine learning. Centralized crowdsourcing platforms face challenges of a single point of failure and a lack of service transparency. Existing blockchain-based crowdsourcing approaches overlook the low scalability problem of permissionless blockchains or inconveniently rely on existing ground-truth data as the root of trust to evaluate quality of workers' answers. We propose a blockchain-based crowdsourcing scheme for ensuring dual fairness (i.e., preventing false-reporting and free-riding) and improving on-chain efficiency concerning on-chain storage and smart contract computation. The proposed scheme does not rely on trusted authorities but rather depends on a public blockchain to guarantee the dual fairness. An efficient and publicly verifiable truth discovery scheme is designed based on majority voting and cryptographic accumulators. This truth discovery scheme aims at inferring ground truth from workers' answers. The ground truth is further utilized to estimate the quality of workers' answers. Additionally, a novel blockchain-based protocol is designed to further reduce on-chain costs while ensuring truthfulness. The scheme has O(n) complexity for both on-chain storage and smart contract computation, regardless of the number of questions, where n denotes the number of workers. Formal security analysis is provided, and extensive experiments are conducted to evaluate effectiveness and performance.

Machine learning (ML) has led to disruptive innovations in many fields, such as medical diagnoses. A key enabler for ML is large training data, but existing data, such as medical data, are not fully exploited by ML because of data silos and privacy concerns. Federated learning (FL) is a promising distributed learning paradigm to address this problem. On the other hand, existing FL approaches are vulnerable to poisoning attacks or privacy leakage from a malicious aggregator or client. This article proposes an auditable FL scheme with Byzantine robustness against the aggregator and client: The aggregator is malicious but available, and the client could perform poisoning attacks. First, the Pedersen commitment scheme (PCS) for homomorphic encryption was applied to preserve privacy and for commitments to the FL process to achieve auditability. The auditability enables clients to verify the correctness and consistency of the entire FL process and to identify parties that misbehave. Second, an efficient technique of divide and conquer was designed based on PCS to allow parties to cooperate and securely aggregate gradients to defend against poisoning attacks. This technique enables clients to share no common secret key and cooperate to decrypt ciphertext, guaranteeing a client’s privacy even if some other clients are corrupted by adversaries. This technique was optimized to tolerate the dropout of clients. This article reports a formal analysis concerning privacy, efficiency, and auditability against malicious participants. Extensive experiments on various benchmark datasets show that the scheme is robust with high model accuracy against poisoning attacks.

Three-dimensional images are frequently used in medical imaging research for classification, segmentation, and detection. However, the limited availability of 3D images hinders research progress due to network training difficulties. Generative methods have been proposed to create medical images using AI techniques. Nevertheless, 2D approaches have difficulty dealing with 3D anatomical structures, which can result in discontinuities between slices. To mitigate these discontinuities, several 3D generative networks have been proposed. However, the scarcity of available 3D images makes training these networks with limited samples inadequate for producing high-fidelity 3D images. We propose a data-guided generative adversarial network to provide high fidelity in 3D image generation. The generator creates fake images with noise using reference code obtained by extracting features from real images. The generator also creates decoded images using reference code without noise. These decoded images are compared to the real images to evaluate fidelity in the reference code. This generation process can create high-fidelity 3D images from only a small amount of real training data. Additionally, our method employs three types of discriminator: volume (evaluates all the slices), slab (evaluates a set of consecutive slices), and slice (evaluates randomly selected slices). The proposed discriminator enhances fidelity by differentiating between real and fake images based on detailed characteristics. Results from our method are compared with existing methods by using quantitative analysis such as Fréchet inception distance and maximum mean discrepancy. The results demonstrate that our method produces more realistic 3D images than existing methods.

Accurately segmenting polyps from colonoscopy images is essential for diagnosing colorectal cancer. Despite the tremendous success of the deep convolutional neural networks in automatic polyp segmentation, it suffers from domain shift issues, where the trained model yields performance deterioration on unseen test datasets. This paper proposes an illumination enhancement-based domain generalization approach to improve the generalization capability of the model on unseen test datasets and alleviate this issue. In particular, an image decomposition module (IDM) was developed to separate colonoscopy images into reflectance, local, and global illumination components. An illumination transform module (ITM) was proposed to augment images with different global illuminations by synthesizing target-like global illumination maps. A novel illumination variance insensitiveness (IViSen) is also introduced to evaluate the robustness of the model against illumination disturbance. IViSen is easy to compute and correlates well with model generalizability. The segmentation performance of the proposed model on four colonoscopy datasets was examined: CVC-ClinicDB, CVC-ColonDB, ETIS-Larib, and Kvasir-SEG. The method outperformed the competitive methods when tested on unseen domains. In particular, the proposed approach yielded 60.82% and 53.19% in terms of mean Dice and IoU, respectively, with 2.06% and 2.31% improvements.

Abstract We consider medical image transformation problems where a grayscale image is transformed into a color image. The colorized medical image should have the same features as the input image because extra synthesized features can increase the possibility of diagnostic errors. In this paper, to secure colorized medical images and improve the quality of synthesized images, as well as to leverage unpaired training image data, a colorization network is proposed based on the cycle generative adversarial network (CycleGAN) model, combining a perceptual loss function and a total variation (TV) loss function. Visual comparisons and experimental indicators from the NRMSE, PSNR, and SSIM metrics are used to evaluate the performance of the proposed method. The experimental results show that GAN-based style conversion can be applied to colorization of medical images. As well, the introduction of perceptual loss and TV loss can improve the quality of images produced as a result of colorization better than the result generated by only using the CycleGAN model.

Crowdsourcing for Human Intelligence Tasks (HIT) has been widely used to crowdsource human knowledge, such as image annotation for machine learning. We use a public blockchain to play the role of traditional centralized HIT systems, such that the blockchain deals with cryptocurrency payments and acts as a trustworthy judge to resolve disputes between a worker and a requester in a decentralized setting, preventing false-reporting and free-riding. Our approach neither uses expensive cryptographic tools, such as zero-knowledge proofs, nor sends the worker's answers to the blockchain. Compared with prior works, our approach significantly reduces on-chain cost: it only requires O(1) on-chain storage and O(log N ) smart contract computation, where N is the question number of a HIT. Additionally, our approach uses known answers or gold standards to determine the worker's answer quality. To motivate the requester to use honest known answers, the requester cannot learn the worker's answers if the answer quality does not meet the requirement. We further provide formal security definitions for our decentralized HIT and prove security of our construction.