Ensuring the structural continuity and completeness of road networks in high-resolution remote sensing imagery remains a major challenge for current deep learning methods, especially under conditions of occlusion caused by vegetation, buildings, or shadows. To address this, we propose a novel post-processing enhancement framework that improves the connectivity and accuracy of initial road extraction results produced by any segmentation model. The method employs a dual-stream encoder architecture, which jointly processes RGB images and preliminary road masks to obtain complementary spatial and semantic information. A core component is the MAF (Multi-Modal Attention Fusion) module, designed to capture fine-grained, long-range, and cross-scale dependencies between image and mask features. This fusion leads to the restoration of fragmented road segments, the suppression of noise, and overall improvement in road completeness. Experiments on benchmark datasets (DeepGlobe and Massachusetts) demonstrate substantial gains in precision, recall, F1-score, and mIoU, confirming the framework’s effectiveness and generalization ability in real-world scenarios.

While collecting training data, even with the manual verification of experts from crowdsourcing platforms, eliminating incorrect annotations (noisy labels) completely is difficult and expensive. In dealing with datasets that contain noisy labels, over-parameterized deep neural networks (DNNs) tend to overfit, leading to poor generalization and classification performance. As a result, noisy label learning (NLL) has received significant attention in recent years. Existing research shows that although DNNs eventually fit all training data, they first prioritize fitting clean samples, then gradually overfit to noisy samples. Mainstream methods utilize this characteristic to divide training data but face two issues: class imbalance in the segmented data subsets and the optimization conflict between unsupervised contrastive representation learning and supervised learning. To address these issues, we propose a Balanced Partitioning and Training framework with Pseudo-Label Relaxed contrastive loss called BPT-PLR, which includes two crucial processes: a balanced partitioning process with a two-dimensional Gaussian mixture model (BP-GMM) and a semi-supervised oversampling training process with a pseudo-label relaxed contrastive loss (SSO-PLR). The former utilizes both semantic feature information and model prediction results to identify noisy labels, introducing a balancing strategy to maintain class balance in the divided subsets as much as possible. The latter adopts the latest pseudo-label relaxed contrastive loss to replace unsupervised contrastive loss, reducing optimization conflicts between semi-supervised and unsupervised contrastive losses to improve performance. We validate the effectiveness of BPT-PLR on four benchmark datasets in the NLL field: CIFAR-10/100, Animal-10N, and Clothing1M. Extensive experiments comparing with state-of-the-art methods demonstrate that BPT-PLR can achieve optimal or near-optimal performance.

Integrating machine learning (ML) with Statistical Process Control (SPC) is important for Industry 4.0 environments. Contemporary manufacturing data exhibit high-dimensionality, autocorrelation, non-stationarity, and class imbalance, which challenge classical SPC assumptions. This systematic review, conducted following the PRISMA 2020 guidelines, provides a problem-driven synthesis that links these data challenges to corresponding methodological families in ML-based SPC. Specifically, we review approaches for (1) high-dimensional and redundant data (dimensionality reduction and feature selection), (2) autocorrelated and dynamic processes (time-series and state-space models), and (3) data scarcity and imbalance (cost-sensitive learning, generative modeling, and transfer learning). Nonlinearity is treated as a cross-cutting property within each category. For each, we outline the mathematical rationale of representative algorithms and illustrate their use with industrial examples. We also summarize open issues in interpretability, thresholding, and real-time deployment. This review offers structured guidance for selecting ML techniques suited to complex manufacturing data and for designing reliable online monitoring pipelines.

High-quality processes, characterized by low defect rates, typically exhibit an exponential distribution for time-between-events (TBE) data. To effectively monitor TBE data, one-sided exponential Adaptive Exponentially Weighted Moving Average (AEWMA) schemes are introduced. However, the run length (RL) distribution varies with the magnitude of the process mean shift, rendering the median run length (MRL) a more pertinent performance metric. This paper investigates the RL properties of one-sided exponential AEWMA schemes using a Markov chain method. An optimal design procedure based on MRL is developed to enhance the one-sided exponential AEWMA scheme. Comparative analyses reveal that the one-sided exponential AEWMA scheme provides better balanced protection against both minor and major shifts in the process mean compared to EWMA-type and Shewhart schemes. Finally, two practical case studies illustrate the application of AEWMA schemes in manufacturing.

Learning with noisy labels (LNL) remains challenging, especially when the identification of clean samples relies heavily on the predictions of the model being trained. In such cases, early-stage selection errors may be reinforced during iterative optimization, leading to unreliable supervision. To alleviate this issue, a two-stage framework, termed CANNE, is proposed by combining Contrastive Language–Image Pre-training (CLIP)-based conservative offline cleaning with Adaptive Nearest Neighbors and eigenvector-based sample selection (ANNE)-based online refinement. Specifically, a high-confidence clean seed set is first constructed using two complementary probability sources derived from frozen CLIP representations and reliability criteria, including class-wise loss modeling and prediction consistency. This seed set is then used as a set of reliable anchors during the subsequent ANNE training process, where online feature- and neighborhood-based refinement further recovers and adjusts sample partitions. In this way, CANNE uses external vision–language priors to provide conservative and persistent guidance while preserving the adaptive recovery ability of online noisy-label learning. Experimental results on CIFAR-10, CIFAR-100, Animal-10N, and Mini-WebVision, together with additional evaluation under open-set noise, show that the proposed method achieves competitive performance across diverse noisy-label settings. In particular, CANNE achieves 96.6% and 96.3% best accuracies on CIFAR-10 under 80% and 90% symmetric noise, respectively, and 81.0% and 79.0% on CIFAR-100 under 20% and 50% symmetric noise. Additional repeated-run, threshold-sensitivity, and runtime analyses further indicate that the CLIP-based seed set provides stable guidance with only moderate computational overhead.

Crop diseases and pests severely threaten global food security. While deep learning has shown promise in controlled settings, real-world field conditions—characterized by complex backgrounds, variable lesion scales, and high inter-class similarity—remain challenging. To overcome limitations in feature representation, scale adaptability, and model efficiency, we propose DPMFNet, a lightweight dual-path network integrating Spatial–Channel Dual-Attention (SCDA) and Multi-Scale Depthwise Separable Convolution (MDSC) modules. SCDA enhances critical regions via dynamic channel–spatial weighting with minimal overhead, while MDSC captures multi-scale contextual information through parallel dilated convolutions. Both are embedded into an improved residual block (AttMDSCBlock) to boost representational power while reducing parameters. A cross-attention mechanism fuses local details and global context from dual pathways, and a lightweight pyramid strategy adaptively integrates features across resolutions. Evaluated on PlantVillage and the AI Challenger 2018 dataset, DPMFNet achieves state-of-the-art accuracy among lightweight models, with only 14.24M parameters and 2.55G FLOPs. It demonstrates superior robustness in complex agricultural environments, balancing performance, efficiency, and deployability.