Search Results (3,911)

With the rapid advancement of autonomous driving technology, Traffic Sign Detection and Recognition (TSDR) has become a critical component for ensuring vehicle safety. However, existing TSDR systems still face significant challenges in accurately detecting partially occluded traffic signs, which poses a substantial risk in real-world applications. To address this issue, this study proposes a comprehensive solution from three perspectives: data augmentation, model architecture enhancement, and dataset construction. We propose an innovative network framework tailored for occluded traffic sign detection. The framework enhances feature representation through a dual-path convolutional mechanism (DualConv) that preserves information flow even when parts of the sign are blocked, and employs a spatial attention module (SEAM) that helps the model focus on visible sign regions while ignoring occluded areas. Finally, we construct the Jinzhou Traffic Sign (JZTS) occlusion dataset to provide targeted training and evaluation samples. Extensive experiments on the public Tsinghua-Tencent 100K (TT-100K) dataset and our JZTS dataset demonstrate the superior performance and strong generalisation capability of our model under occlusion conditions. This work not only advances the robustness of TSDR systems for autonomous driving but also provides a valuable benchmark for future research. Full article

The rapid expansion of Internet of Things (IoT)-enabled smart grids has intensified the need for reliable fault detection and autonomous self-healing under non-stationary operating conditions characterized by frequent concept drift. To address the limitations of static and single-strategy adaptive models, this paper proposes H-CLAS, a novel Hybrid Continual Learning for Adaptive Self-healing framework that unifies regularization-based, memory-based, architectural, and meta-learning strategies within a single adaptive pipeline. The framework integrates convolutional neural networks (CNNs) for fault detection, graph neural networks for topology-aware fault localization, reinforcement learning for self-healing control, and a hybrid drift detection mechanism combining ADWIN and Page–Hinkley tests. Continual adaptation is achieved through the synergistic use of Elastic Weight Consolidation, memory-augmented replay, progressive neural network expansion, and Model-Agnostic Meta-Learning for rapid adaptation to emerging drifts. Extensive experiments conducted on the Smart City Air Quality and Network Intrusion Detection Dataset (NSL-KDD) demonstrate that H-CLAS achieves accuracy improvements of 12–15% over baseline methods, reduces false positives by over 50%, and enables 2–3× faster recovery after drift events. By enhancing resilience, reliability, and autonomy in critical IoT-driven infrastructures, the proposed framework contributes to improved grid stability, reduced downtime, and safer, more sustainable energy and urban monitoring systems, thereby providing significant societal and environmental benefits. Full article

Credit card fraud detection remains a major challenge due to extreme class imbalance and evolving attack patterns. This paper proposes a practical hybrid pipeline that combines conditional tabular generative adversarial networks (CTGANs) for targeted minority-class synthesis with Light Gradient Boosting Machine (LightGBM) for classification. Inspired by symmetry principles in machine learning, we leverage the adversarial equilibrium of CTGAN to generate realistic fraudulent transactions that maintain distributional symmetry with real fraud patterns, thereby preserving the structural and statistical balance of the original dataset. Synthetic fraud samples are merged with real data to form augmented training sets that restore the symmetry of class representation. We evaluate Simple Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU) classifiers, and a LightGBM model on a public dataset using stratified 5-fold validation and an independent hold-out test set. Models are compared using sensitivity, precision, F-measure(F₁), and area under the precision–recall curve (PR-AUC), which reflects symmetry between detection and false-alarm trade-offs. Results show that CTGAN-based augmentation yields large and consistent gains across architectures. The best-performing configuration, CTGAN + LightGBM, attains sensitivity = 0.986, precision = 0.982, F₁ = 0.984, and PR-AUC = 0.918 on the test data, substantially outperforming non-augmented baselines and recent methods. These findings indicate that conditional synthetic augmentation materially improves the detection of rare fraud modes while preserving low false-alarm rates, demonstrating the value of symmetry-aware data synthesis in classification under imbalance. We discuss generation-quality checks, risk of distributional shift, and deployment considerations. Future work will explore alternative generative models with explicit symmetry constraints and time-aware production evaluation. Full article

Continuous missing data is a prevalent challenge in long-term air quality monitoring, undermining the reliability of public health protection and sustainable urban development. In this paper, we propose ConFill, a novel contrastive learning-based framework for reconstructing continuous missing data in air quality time series. By leveraging temporal continuity as a supervisory signal, our method constructs positive sample pairs from adjacent subsequences and negative pairs from distant and shuffled segments. Through contrastive learning, the model learns robust representations that preserve intrinsic temporal dynamics, and enable accurate imputation of continuous missing segments. A novel data augmentation strategy is proposed, to integrate noise injection, subsequence masking, and time warping to enhance the diversity and representativeness of training samples. Extensive experiments are conducted on a large scale real-world dataset comprising multi-pollutant observations from 209 monitoring stations across China over a three-year period. Results show that ConFill outperforms baseline imputation methods under various missing scenarios, especially in reconstructing long consecutive gaps. Ablation studies confirm the effectiveness of both the contrastive learning module and the proposed augmentation technique. Full article

In the field of Natural Language Inference (NLI), model interpretability remains an urgent and unresolved challenge. Existing interpretability-oriented annotated datasets are highly limited, and manually constructing natural language explanations is both costly and inconsistent, making it difficult to balance model performance and interpretability. To address this issue, this paper proposes an interpretable NLI framework based on active learning, Explanation Generation Model-Prediction Model (EGM-PM), and designs an active learning sampling algorithm, Explanation-aware Transition from Clustering to Margin (ETCM), that incorporates natural-language explanation information. In this framework, Large Language Models (LLMs) are employed to automate explanation annotation, reducing dependence on human experts in traditional active learning. A small number of high-value samples obtained via ETCM sampling are used to train the EGM, whose generated natural-language explanations are then used to guide the PM in label inference. Experimental results show that data sampled by ETCM substantially enhance the model’s ability to learn relational and logical structures between premise–hypothesis pairs. Compared with other active learning algorithms, ETCM approaches full-data performance more rapidly while using significantly fewer labeled samples. This finding confirms the value of natural language explanation semantics in improving both model performance and interpretability. Furthermore, this paper employs prompt engineering to construct an interpretability-oriented NLI dataset, Explainable Natural Language Inference (ExNLI), which augments traditional premise–hypothesis pairs with natural-language explanations. Human and automated evaluations confirm the consistency and faithfulness of these explanations. The dataset has been publicly released, offering a low-cost and scalable data construction approach for future research on explainable NLI. Full article

This work addresses the challenges of inadequate data acquisition and the limited availability of labeled samples for shearer rocker arm bearing faults by developing a data augmentation methodology that synergistically incorporates the Genetic Algorithm-optimized Wavelet Transform Symmetrical Dot Pattern (GA-WT-SDP) with a Wasserstein Conditional Generative Adversarial Network (WCGAN). In the initial step, the Genetic Algorithm (GA) is employed to refine the mapping parameters of the Wavelet Transform Symmetrical Dot Pattern (WT-SDP), facilitating the transformation of raw vibration signals into advanced and discriminative graphical representations. Thereafter, the Wasserstein distance in conjunction with a gradient penalty mechanism is introduced through the WCGAN, thereby ensuring higher-quality generated samples and improved stability during model training. Experimental results validate that the proposed approach yields accelerated convergence and superior performance in sample generation. The augmented data significantly bolsters the generalization ability and predictive accuracy of fault diagnosis models trained on small datasets, with notable gains achieved in deep architectures (CNNs, LSTMs). The research substantiates that this technique helps overcome overfitting, enhances feature representation capacity, and ensures consistently high identification accuracy even in complex working environments. Full article

As a globally significant food crop, the assessment of wheat quality is essential for ensuring food security and enhancing the processing quality of agricultural products. Conventional methods for assessing wheat kernel quality are often inefficient and markedly subjective, which hampers their ability to accurately distinguish the complex and diverse phenotypic characteristics of wheat kernels. To tackle the aforementioned issues, this study presents an enhanced recognition method for defective wheat kernels, based on the EfficientNet-B1 architecture. Building upon the original EfficientNet-B1 network structure, this approach incorporates the lightweight attention mechanism known as CBAM (Convolutional Block Attention Module) to augment the model’s capacity to discern features in critical regions. Simultaneously, it modifies the classification head structure to facilitate better alignment with the data, thereby enhancing accuracy. The experiment employs a self-constructed dataset comprising five categories of wheat kernels—perfect wheat kernels, insect-damaged wheat kernels, scab-damaged wheat kernels, moldy wheat kernels, and black germ wheat kernels—which are utilized for training and validation purposes. The results indicate that the enhanced model attains a classification accuracy of 99.80% on the test set, reflecting an increase of 2.6% compared to its performance prior to the enhancement. Furthermore, the Precision, Recall, and F1-score all demonstrated significant improvements. The proposed model achieves near-perfect performance on several categories under controlled experimental conditions, with particularly high precision and recall for scab-damaged and insect-damaged kernels. This study demonstrates the efficacy of the enhanced EfficientNet-B1 model in the recognition of defective wheat kernels and offers novel technical insights and methodological references for intelligent wheat quality assessment. Full article

In the era of the digital food economy, high-fidelity food images are critical for applications ranging from visual e-commerce presentation to automated dietary assessment. However, developing robust computer vision systems for food analysis is often hindered by data scarcity for long-tail or regional dishes. To address this challenge, we propose a novel high-fidelity food image synthesis framework as an effective data augmentation tool. Unlike generic generative models, our method introduces an Ingredient-Aware Diffusion Model based on the Masked Diffusion Transformer (MaskDiT) architecture. Specifically, we design a Label and Ingredients Encoding (LIE) module and a Cross-Attention (CA) mechanism to explicitly model the relationship between food composition and visual appearance, simulating the “cooking” process digitally. Furthermore, to stabilize training on limited data samples, we incorporate a linear interpolation strategy into the diffusion process. Extensive experiments on the Food-101 and VireoFood-172 datasets demonstrate that our method achieves state-of-the-art generation quality even in data-scarce scenarios. Crucially, we validate the practical utility of our synthetic images: utilizing them for data augmentation improved the accuracy of downstream food classification tasks from 95.65% to 96.20%. This study provides a cost-effective solution for generating diverse, controllable, and realistic food data to advance smart food systems. Full article

This systematic review examines 205 studies on the use of Large Language Models (LLMs) in psychiatry, psychology, psychotherapy, and clinical workflows. Furthermore, studies that directly evaluated at least one LLM in a mental health context were included in the extended detailed analysis. GPT-4 and GPT-3.5 were the most commonly assessed models. Although LLMs showed promising short-term performance across domains, most evaluations relied on small, non-longitudinal datasets and single-session testing, limiting generalizability. The evidence indicates rapid growth but significant methodological inconsistency, emphasizing the need for more diverse datasets, standardized evaluation, and long-term validation before clinical integration. This review also examines how LLMs are being incorporated into mental health practice, outlining key challenges, limitations, and emerging opportunities. Ethical, clinical, and technological considerations are proposed to guide responsible adoption. Given the complexity of mental health care, a multidisciplinary, human-centered approach remains essential to ensure that future LLM applications augment—rather than replace—professional expertise. Full article

Cost code verification in state-funded construction projects remains a labor-intensive and error-prone task, particularly given the structural heterogeneity of project estimates and the prevalence of malformed codes, inconsistent units of measurement (UoMs), and locally modified price components. This study evaluates a deterministic GPT-based assistant designed to automate Vietnam’s regulatory verification. The assistant was developed and iteratively refined across four Action Research cycles. Also, the system enforces strict rule sequencing and dataset grounding via Python-governed computations. Rather than relying on probabilistic or semantic reasoning, the system performs strictly deterministic checks on code validity, UoM alignment, and unit price conformity in material (MTR), labor (LBR), and machinery (MCR), given the provincial unit price books (UPBs). Deterministic equality is evaluated either on raw numerical values or on values transformed through explicitly declared, rule-governed operations, preserving auditability without introducing tolerance-based or inferential reasoning. A dedicated exact-match mechanism, which is activated only when a code is invalid, enables the recovery of typographical errors only when a project item’s full price vector well matches a normative entry. Using twenty real construction estimates (16,100 rows) and twelve controlled error-injection cases, the study demonstrates that the assistant executes verification steps with high reliability across diverse spreadsheet structures, avoiding ambiguity and maintaining full auditability. Deterministic extraction and normalization routines facilitate robust handling of displaced headers, merged cells, and non-standard labeling, while structured reporting provides line-by-line traceability aligned with professional verification workflows. Practitioner feedback confirms that the system reduces manual tracing effort, improves evaluation consistency, and supports documentation compliance with human judgment. This research contributes a framework for large language model (LLM)-orchestrated verification, demonstrating how Action Research can align AI tools with domain expectations. Furthermore, it establishes a methodology for deploying LLMs in safety-critical and regulation-driven environments. Limitations—including narrow diagnostic scope, unlisted quotation exclusion, single-province UPB compliance, and sensitivity to extreme spreadsheet irregularities—define directions for future deterministic extensions. Overall, the findings illustrate how tightly constrained LLM configurations can augment, rather than replace, professional cost verification practices in public-sector construction. Full article

Tomato leaf diseases cause substantial yield and quality losses worldwide, yet reliable detection in real fields remains challenging. Two practical bottlenecks dominate current research: (i) limited data, including small samples for rare diseases, class imbalance, and noisy field images, and (ii) multimodal heterogeneity, where RGB images, textual symptom descriptions, spectral cues, and optional molecular assays provide complementary but hard-to-align evidence. This review summarizes recent advances in tomato leaf disease detection under these constraints. We first formalize the problem settings of limited and multimodal data and analyze their impacts on model generalization. We then survey representative solutions for limited data (transfer learning, data augmentation, few-/zero-shot learning, self-supervised learning, and knowledge distillation) and multimodal fusion (feature-, decision-, and hybrid-level strategies, with attention-based alignment). Typical model–dataset pairs are compared, with emphasis on cross-domain robustness and deployment cost. Finally, we outline open challenges—including weak generalization in complex field environments, limited interpretability of multimodal models, and the absence of unified multimodal benchmarks—and discuss future opportunities toward lightweight, edge-ready, and scalable multimodal systems for precision agriculture. Full article

Surface reconstruction from 3D point clouds has a wide range of applications. In this paper, we focus on the reconstruction from raw, sparse point clouds. Although some existing methods work on this topic, the results often suffer from geometric defects. To solve this problem, we propose a novel method that optimizes a neural network (referred to as signed distance function) to fit the Signed Distance Field (SDF) from sparse point clouds. The signed distance function is optimized by projecting query points to its iso-surface accordingly. Our key idea is to encourage both the direction and distance of projection to be correct through the supervision provided by a nearest neighbor prior. In addition, we mitigate the error propagated from the prior function by augmenting the low-frequency components in the input. In our implementation, the nearest neighbor prior is trained with a large-scale local geometry dataset, and the positional encoding with a specified spectrum is used as a regularization for the optimization process. Experiments on the ShapeNetCore dataset demonstrate that our method achieves better accuracy than SDF-based methods while preserving smoothness. Full article

Aiming at the insufficient capture of temporal dependence and weak coupling of external factors in multivariate load forecasting, this paper proposes a Transformer model integrating timestamp-based positional embedding and convolution-augmented attention. The model enhances temporal modeling capability through timestamp-based positional embedding, optimizes local contextual representation via convolution-augmented attention, and achieves deep fusion of load data with external factors such as temperature, humidity, and electricity price. Experiments based on the 2018 full-year load dataset for a German region show that the proposed model outperforms single-factor and multi-factor LSTMs in both short-term (24 h) and long-term (cross-month) forecasting. The research results verify the model’s accuracy and stability in multivariate load forecasting, providing technical support for smart grid load dispatching. Full article

Accurate downhole depth measurement is essential for oil and gas well operations, directly influencing reservoir contact, production efficiency, and operational safety. Collar correlation using a casing collar locator (CCL) is fundamental for precise depth calibration. While neural network has achieved significant progress in collar recognition, preprocessing methods for such applications remain underdeveloped. Moreover, the limited availability of real well data poses substantial challenges for training neural network models that require extensive datasets. This paper presents a system integrated into a downhole toolstring for CCL log acquisition to facilitate dataset construction. Comprehensive preprocessing methods for data augmentation are proposed, and their effectiveness is evaluated using baseline neural network models. Through systematic experimentation across diverse configurations, the contribution of each augmentation method is analyzed. Results demonstrate that standardization, label distribution smoothing (LDS), and random cropping are fundamental prerequisites for model training, while label smoothing regularization (LSR), time scaling, and multiple sampling significantly enhance model generalization capabilities. Incorporating the proposed augmentation methods into the two baseline models results in maximum F1 score improvements of 0.027 and 0.024 for the TAN and MAN models, respectively. Furthermore, applying these techniques yields F1 score gains of up to 0.045 for the TAN model and 0.057 for the MAN model compared to prior studies. Performance evaluation on real CCL waveforms confirms the effectiveness and practical applicability of our approach. This work addresses the existing gaps in data augmentation methodologies for training casing collar recognition models under CCL data-limited conditions, and provides a technical foundation for the future automation of downhole operations. Full article

Precise identification of early peanut leaf spot is strategically significant for safeguarding oilseed supplies and reducing pesticide reliance. However, general-purpose detectors face severe domain adaptation bottlenecks in unstructured field environments due to small feature dissipation, physical occlusion, and class imbalance. To address this, this study constructs a dataset spanning two phenological cycles and proposes POD-YOLO, a physics-aware and dynamics-optimized lightweight framework. Anchored on the YOLOv10n architecture and adhering to a “data-centric” philosophy, the framework optimizes the parameter convergence path via a synergistic “Augmentation-Loss-Optimization” mechanism: (1) Input Stage: A Physical Domain Reconstruction (PDR) module is introduced to simulate physical occlusion, blocking shortcut learning and constructing a robust feature space; (2) Loss Stage: A Loss Manifold Reshaping (LMR) mechanism is established utilizing dual-branch constraints to suppress background gradients and enhance small target localization; and (3) Optimization Stage: A Decoupled Dynamic Scheduling (DDS) strategy is implemented, integrating AdamW with cosine annealing to ensure smooth convergence on small-sample data. Experimental results demonstrate that POD-YOLO achieves a 9.7% precision gain over the baseline and 83.08% recall, all while maintaining a low computational cost of 8.4 GFLOPs. This study validates the feasibility of exploiting the potential of lightweight architectures through optimization dynamics, offering an efficient paradigm for edge-based intelligent plant protection. Full article