Efficient and highly accurate road crack detection algorithms are particularly important in road inspection systems. However, some of their limitations have gradually come to the fore as the target detection aspect has become more in-depth. Existing road target detection algorithms face the difficulty of capturing long-range dependencies, resulting in limited feature expressiveness and high leakage rates in small target detection scenarios ( e.g ., road fine crack identification). Therefore, in this article, we propose an improved model, You Only Look Once (YOLO)-SW, based on YOLO11n. Firstly, based on the structure of the C2f module, we form a deep feature processing chain by stacking n sp modules through nn.ModuleList (the module is responsible for multiscale feature aggregation and attention mechanism), and propose the CSP_PMSFA module, which enhances the ability of small-target detection. The method replaces C3k2 of YOLO11n with the Cross Stage Partial Path Multiscale Spatial Feature Aggregation (CSP_PMSFA) module that enhances the feature expression ability by combining multi-scale feature aggregation and attention mechanism, which effectively improves the feature expression accuracy of small targets. At the same time, the CGAFusion module is added to enhance the feature expression ability by combining spatial attention, channel attention, and pixel attention. Experiments on the pavement crack detection Computer Vision Project dataset show that the mean average accuracy of the improved YOLO-SW model (mean average precision at Intersection of Union 0.5 (mAP@ 0.5)) reaches 58.2%, which is an improvement of 8.7 percentage points over the baseline model YOLO11n. The experimental results validate the significant advantages of YOLO-SW for crack detection in complex road scenarios.

Secure data collaboration among mutually distrustful organizations requires more than encrypted storage: it also needs accountable ownership control, auditable access governance, privacy-preserving transaction execution, and reliable settlement when data are exchanged as digital assets. This paper proposes TrustTrade, a unified multi-party secure data management and transaction framework designed for cross-organization data sharing, trading, and compliance-sensitive analytics. TrustTrade integrates policy-bound data capsules, a tamper-evident provenance ledger, adaptive threshold escrow, verifiable data-payment settlement, and selective audit with revocation rebinding. On four real-dataset-derived workloads, TrustTrade reaches a $90.4$–$94.8\%$ settlement rate, with a $92.5\%$ average that is $6.4$ percentage points higher than the strongest baseline average. Under adversarial request injection, TrustTrade reduces unauthorized release to $0.31\%$ and atomicity violation to $0.38\%$, corresponding to $93.6\%$ and $93.0\%$ reductions compared with Plain-Market, respectively; compared with Fixed-Escrow, unauthorized release is reduced by $77.4\%$. TrustTrade also achieves $96.7\%$ dispute-resolution accuracy while maintaining practical settlement latency. These results indicate that jointly designing secure data management and secure data transaction protocols offers a practical path toward trustworthy multi-party data ecosystems.

Recognizing visual artworks is difficult because images of art combine large within-class diversity, ambiguous boundaries, and non-photographic textures with intertwined cues related to the artist, genre, style, and historical period. Standard convolutional or transformer models are commonly inherited from natural-image classification, and such backbones may fail to emphasize the visual statistics that distinguish paintings, drawings, prints, and other artistic media. This paper presents ArtNAS-Style, a differentiable neural architecture search framework that embeds artistic style priors into the architecture search stage for visual art recognition. Rather than optimizing architecture choices only through classification loss, the framework employs a style-prior-aware bilevel objective that jointly accounts for recognition accuracy, diversity of style-sensitive features, preservation of brushstroke-like texture, and robustness across stylistic domains. The proposed Style Prior Encoding Module derives Gram-statistic, spectral, and palette-distribution cues from each artwork, and the Prior-Conditioned Search Cell uses these cues to adaptively modulate candidate operations during the search. A Style-Balanced Architecture Regularizer further discourages the discovered architecture from specializing to dominant artistic categories and promotes discriminative features across heterogeneous art domains. The search model is compact, trainable end to end, and transferable to multiple recognition tasks. Our evaluation protocol covers style, artist, school, and object-type recognition, in addition to parameter-matched ablations, robustness, statistical testing, and computational costs.

Open-world object detection requires knowledge of categories and discovery of new objects never seen in training. Full model fine tuning fails often in such dynamic environment. Fully tuned models overfit known classes and lose their sensitivity to unknown objects at high computational cost. To solve this problem, we propose a Parameter-Efficient Open-World Object Detection (PE-OWOD) framework, a simple and light approach to retrain. Instead of updating the whole network, we lock backbone and encoder to maintain stable visual priors and inject compact Residual adapters only into decoder to adapt tasks. We also introduce VOS, which defines explicit decision boundary for open space with optional semantic initialization. MS-COCO benchmarks show remarkable efficiency advantages: Update less than 27% of models, PE-OWOD achieves 64.7% Unknown Recall (significantly outperform fully tuned baselines), and GPU memory usage is reduced by 86%. These results indicate that effective parameter adaptation is not a constraint; rather, it is a reliable and robust open-world detection strategy.

Federated learning (FL) enables collaborative model training without centralizing raw data, yet practical deployments remain vulnerable to security threats such as Byzantine model poisoning, stealthy backdoor implantation, and integrity attacks that exploit the opacity of client updates. This paper presents SecureFedGuard, a security-centric FL framework that introduces a novel combination of (i) dual-view update authentication that binds each client update to a lightweight stochastic gradient fingerprint, enabling server-side integrity screening without accessing client data, and (ii) backdoor-resilient aggregation driven by cross-round spectral forensics and adaptive coordinate-wise trimming guided by an estimated benign subspace. SecureFedGuard is designed to be compatible with secure aggregation and does not require trusted hardware, public datasets for pretraining, or expensive per-client verification. We provide a simple robustness analysis that clarifies when benign updates dominate the estimated subspace under mixed benign/malicious participation. Experiments on real FL benchmarks (vision and language) under diverse threat models show that SecureFedGuard substantially improves clean accuracy and backdoor attack success rate compared with strong baselines, while adding modest communication and computation overhead. These results suggest a practical path toward integrity-preserving and backdoor-resistant FL without weakening the privacy boundary between clients and the server.

Open-world object detection requires knowledge of categories and discovery of new objects never seen in training. Full model fine tuning fails often in such dynamic environment. Fully tuned models overfit known classes and lose their sensitivity to unknown objects at high computational cost. To solve this problem, we propose PE-OWOD, a simple and light approach to retrain. Instead of updating the whole network, we lock backbone and encoder to maintain stable visual priors and inject compact Residual adapters only into decoder to adapt tasks. We also introduce VOS, which defines explicit decision boundary for open space with optional semantic initialization. MS-COCO benchmarks show remarkable efficiency advantages: Update less than 27% of models, PE-OWOD achieves 64.7% Unknown Recall (significantly outperform fully tuned baselines), and GPU memory usage is reduced by 86%. These results indicate that effective parameter adaptation is not a constraint; rather, it is a reliable and robust open-world detection strategy.