Search Results (2,048)

High-quality image restoration (IR) is a fundamental task in computer vision, aiming to recover a clear image from its degraded version. Prevailing methods typically employ a static inference pipeline, neglecting the spatial variability of image content and degradation, which makes it difficult for them to adaptively handle complex and diverse restoration scenarios. To address this issue, we propose a novel adaptive image restoration framework named Hybrid Convolutional Transformer with Dynamic Prompting (HCTDP). Our approach introduces two key architectural innovations: a Spatially Aware Dynamic Prompt Head Attention (SADPHA) module, which performs fine-grained local restoration by generating spatially variant prompts through real-time analysis of image content and a Gated Skip-Connection (GSC) module that refines multi-scale feature flow using efficient channel attention. To guide the network in generating more visually plausible results, the framework is optimized with a hybrid objective function that combines a pixel-wise L1 loss and a feature-level perceptual loss. Extensive experiments on multiple public benchmarks, including image deraining, dehazing, and denoising, demonstrate that our proposed HCTDP exhibits superior performance in both quantitative and qualitative evaluations, validating the effectiveness of the adaptive restoration framework while utilizing fewer parameters than key competitors. Full article

Fire hazards during cotton picker operations pose a significant safety concern, primarily caused by cotton blockages and friction-induced heat generation between the picking spindle and seed cotton under high-load conditions. Existing fire monitoring systems typically employ a uniform temperature threshold across multiple sensors. However, this approach overlooks the distinct characteristics of different cotton picker mechanisms and the influence of seed cotton moisture content, resulting in frequent false alarms and missed detections. To address these issues, this study pioneers and tests a dynamic, tiered temperature threshold warning strategy. This approach accounts for key cotton picker components and varying seed cotton moisture content (MC), specifically MC 9–12% and MC 12–15%. Additionally, based on the operational characteristics of the cotton conveying tube, this study proposes monitoring the wall surface temperature of the conveying tube and investigates the threshold for this temperature. Results indicate that during seed cotton open burning, the average temperature is 324 °C for MC < 9%, 261.9 °C for MC 9–12%, and 178.4 °C for MC 12–15%. After transitioning to smoldering, the temperatures were 226.6 °C, 191.5 °C, and 163.5 °C, respectively, with 163.5 °C being the lowest threshold for seed cotton open burning in the cotton bin. For smoldering seed cotton, the temperature thresholds were 240 °C for MC < 9% and MC 9–12%, and 280 °C for MC 12–15%. The temperature threshold for the cotton conveyor pipe wall surface was 49 °C. The friction-induced heat generation temperature threshold at the picking head, determined through combined testing and simulation, is set at 289 °C for MC < 9%, 306 °C for MC 9–12%, and 319 °C for MC 12–15%. The aforementioned tiered early warning strategy, developed through multi-source experiments and simulations, can be directly configured into controllers. It enables dynamic threshold alarms based on harvester location, seed cotton moisture content, and temperature zones, providing quantitative support for cotton harvester fire monitoring and risk management. Full article

Accurate pedestrian detection from a robotic perspective has become increasingly critical, especially in complex environments such as crowded and high-density populations. Existing methods have low accuracy due to multi-scale pedestrians and dense occlusion in complex environments. To address the above drawbacks, a dense pedestrian detection network architecture based on YOLOv8n (DPDN-YOLOv8) was introduced for complex environments. The network aims to improve robots’ pedestrian detection in complex environments. Firstly, the C2f modules in the backbone network are replaced with C2f_ODConv modules integrating omni-dimensional dynamic convolution (ODConv) to enable the model’s multi-dimensional feature focusing on detected targets. Secondly, the up-sampling operator Content-Aware Reassembly of Features (CARAFE) is presented to replace the Up-Sample module to reduce the loss of the up-sampling information. Then, the Adaptive Spatial Feature Fusion detector head with four detector heads (ASFF-4) was introduced to enhance the system’s ability to detect small targets. Finally, to accelerate the convergence of the network, the Focaler-Shape-IoU is utilized to become the bounding box regression loss function. The experimental results show that, compared with YOLOv8n, the mAP@0.5 of DPDN-YOLOv8 increases from 80.5% to 85.6%. Although model parameters increase from $3\times {10}^6$ to $5.2\times {10}^6$, it can still meet requirements for deployment on mobile devices. Full article

Virtual commands serve as the essential framework for establishing interaction between the virtual pilot and the MCP in autopilot scenarios. Conventional proportional-integral-derivative (PID) controllers are insufficient to ensure accurate flight trajectories due to system hysteresis. To overcome this limitation, a quaternary correlation prediction compensation PID (QCPC-PID) approach is introduced for computing virtual heading commands in autopilot tasks. The method integrates multi-feature statistics, entropy-based predictive compensation, and quaternary correlations. First, flight trajectory error statistics are dynamically calculated using signed error distances to assess deviation levels. Second, a predictive structure based on information entropy is applied to enhance PID compensation. Third, quaternary correlation dependence is established to generate virtual heading commands. The findings confirm the effectiveness of the method in improving flight convergence. The incorporation of predictive structures and quaternary correlations is critical for achieving predictive compensation during PID tuning, thereby reducing flight trajectory deviations. The quaternary correlation prediction compensation method ensures superior performance of PID control in modeling heading adjustment behavior under autopilot conditions. Full article

Pipeline leak detection through acoustic signal measurement faces critical challenges, including insufficient utilization of time-frequency domain features, poor adaptability to noisy environments, and inadequate exploitation of frequency-domain prior knowledge in existing deep learning approaches. This paper proposes a Multi-Scale Frequency-Aware Transformer (MSFAT) architecture that integrates measurement-based acoustic signal analysis with artificial intelligence techniques. The MSFAT framework consists of four core components: a frequency-aware embedding layer that achieves joint representation learning of time-frequency dual-domain features through parallel temporal convolution and frequency transformation, a multi-head frequency attention mechanism that dynamically adjusts attention weights based on spectral distribution using frequency features as modulation signals, an adaptive noise filtering module that integrates noise detection, signal enhancement, and adaptive fusion functions through end-to-end joint optimization, and a multi-scale feature aggregation mechanism that extracts discriminative global representations through complementary pooling strategies. The proposed method addresses the fundamental limitations of traditional measurement-based detection systems by incorporating domain-specific prior knowledge into neural network architecture design. Experimental validation demonstrates that MSFAT achieves 97.2% accuracy and an F1-score, representing improvements of 10.5% and 10.9%, respectively, compared to standard Transformer approaches. The model maintains robust detection performance across signal-to-noise ratio conditions ranging from 5 to 30 dB, demonstrating superior adaptability to complex industrial measurement environments. Ablation studies confirm the effectiveness of each innovative module, with frequency-aware mechanisms contributing most significantly to the enhanced measurement precision and reliability in pipeline leak detection applications. Full article

Multi-modal knowledge graph completion (MMKGC) aims to complete knowledge graphs by integrating structural information with multi-modal (e.g., visual, textual, and numerical) features and leveraging cross-modal reasoning within a unified semantic space to infer and supplement missing factual knowledge. Current MMKGC methods have advanced in terms of integrating multi-modal information but have overlooked the imbalance in modality importance for target entities. Treating all modalities equally dilutes critical semantics and amplifies irrelevant information, which in turn limits the semantic understanding and predictive performance of the model. To address these limitations, we proposed a modality information aggregation graph attention network with adversarial training for multi-modal knowledge graph completion (MIAGAT-AT). MIAGAT-AT focuses on hierarchically modeling complex cross-modal interactions. By combining the multi-head attention mechanism with modality-specific projection methods, it precisely captures global semantic dependencies and dynamically adjusts the weight of modality embeddings according to the importance of each modality, thereby optimizing cross-modal information fusion capabilities. Moreover, through the use of random noise and multi-layer residual blocks, the adversarial training generates high-quality multi-modal feature representations, thereby effectively enhancing information from imbalanced modalities. Experimental results demonstrate that our approach significantly outperforms 18 existing baselines and establishes a strong performance baseline across three distinct datasets. Full article

South Africa’s urbanization is often driven by poverty, unemployment, and limited resource access. Unearned income, such as social grants and other sources, has contributed to poverty alleviation. However, concerns have also been raised that this unearned support may reduce individuals’ motivation to pursue earned income opportunities. This study investigates whether a two-step modelling approach provides better insight than a single-framework model to assess the influence of youths’ access to resources on household income generation. The results indicate that the two-step model is more effective, as different factors influence the decision to earn income and the amount earned. Youth unemployment and household receipt of remittances had similar effects on both the decision to earn income and the amount earned. In contrast, youth involvement in agriculture was positively associated with the decision to earn income but negatively associated with the amount of income. Youth-headed households face additional constraints due to limited access to and ownership of productive resources. The study concludes that a two-step approach provides more information and thus a more accurate understanding of rural income dynamics. Enhancing youth access to quality resources and evaluating the effectiveness of support programs are essential for fostering income generation and improving rural livelihoods. Full article

In the global agricultural production system, maintaining and improving soil quality are core elements for ensuring food security and sustainable agricultural development. As a key indicator of soil quality, the content and dynamic change in soil organic carbon have a profound impact on the physical, chemical and biological properties of soil, and play a decisive role in soil fertility, structural stability, water and fertilizer conservation capacity and microbial activity. However, its decomposition is slow, and a large number of straws returning to the field will impact crop growth; its combination with irrigation is a more reasonable solution, as it can significantly improve the soil environment, increase soil moisture and promote straw decomposition. Therefore, in order to further study the effects of different irrigation methods and straw-returning combinations on soil active-carbon content, an experiment was carried out in long-term arid and semi-arid areas under in-field corn cultivation during 2019–2020. Three irrigation modes were designed—flood irrigation (BI), shallow drip irrigation (SD) and drip irrigation under film (DP)—and straw returning (CS) and no straw returning (CK) were set up, with irrigation applied at critical corn growth stages (internode elongation, heading, bell mouth stage) to support plant growth. The results are as follows: (1) The content of soil organic carbon in different treatments had a gradual upward trend with the advance of growth period; the content of soil organic carbon in DP treatment was significantly higher than that in SD and BI treatment under the same straw returning mode, indicating that drip irrigation under film and straw-returning mode can synergistically improve soil fertility and organic carbon content. (2) Different irrigation methods and straw-returning methods have significant effects on the content of soil active organic carbon components. Different drip irrigation modes can significantly improve the content of soil POC and MBC compared with flood irrigation. The Kos of SD treatment is significantly higher than that of other irrigation treatments, and the CPMI is lower than that of the other two irrigation methods, indicating that the soil organic carbon of SD treatment is more stable. Therefore, under straw-returning conditions, drip irrigation can significantly improve the carbon content of soil components and the management index of soil carbon pool, thus significantly increasing the accumulation of soil organic matter. This study discussed the effects of straw returning on soil organic carbon composition and soil carbon pool index under different irrigation methods to provide theoretical and practical bases for the selection and promotion of straw-returning methods and rational irrigation methods in semi-arid areas. Full article

Background and objectives: Melanocytic nevi are among the most common skin lesions, yet their relationship with the peripheral nervous system has remained understudied. Given the neural crest origin of melanocytes and Schwann cells, and the neurotrophic signaling capabilities of pigment cells, this study aimed to investigate the density of nerve fibers within nevi and assess how it varies with respect to histological subtype and anatomical location. Materials and Methods: A total of 90 nevi were analyzed, including junctional, compound, and intradermal types, distributed across the head, trunk, and limbs. Immunofluorescence staining for the pan-neuronal marker PGP 9.5 and for CGRP were performed and nerve fiber density was quantified. Statistical evaluation using two-way ANOVA revealed that both nevus type and anatomical site significantly influenced the degree of total innervation. Results: Junctional nevi demonstrated the highest total nerve fiber density, significantly exceeding that of compound and intradermal nevi. Likewise, nevi located on the head exhibited a significantly greater density of PGP 9.5-positive nerve fibers compared to those on the trunk and limbs. No significant correlation was observed between nevus type and location, suggesting that both factors contribute independently to the differences in innervation. CGRP-positive innervation was uniform regardless of the histological type of nevus and anatomical location. Conclusions: These findings likely reflect the facts that junctional nevi reside at the dermo-epidermal junction, where nerve fibers are most abundant, while the skin of the head and neck is well known to be more richly innervated than other regions. In contrast, analysis of CGRP-positive fibers suggests that the heterogeneity detected with PGP 9.5 is primarily driven by other neuronal populations. The results support the hypothesis of a dynamic relationship between nevi and the peripheral nervous system, potentially mediated by neurotrophic factors. Understanding this interaction may provide insight into nevus biology, sensory symptoms reported in some lesions, and the evolving role of nerves in the tumor microenvironment. Full article

Greenhouse automation has become increasingly important in facility agriculture, yet multi-span glass greenhouses pose both scientific and practical challenges for autonomous mobile robots. Scientifically, solid-state LiDAR is vulnerable to glass-induced reflections, sparse geometric features, and narrow vertical fields of view, all of which undermine Simultaneous Localization and Mapping (SLAM)-based localization and mapping. Practically, large-scale crop production demands accurate inter-row navigation and efficient rail switching to reduce labor intensity and ensure stable operations. To address these challenges, this study presents an integrated localization-navigation framework for mobile robots in multi-span glass greenhouses. In the intralogistics area, the LiDAR Inertial Odometry-Simultaneous Localization and Mapping (LIO-SAM) pipeline was enhanced with reflection filtering, adaptive feature-extraction thresholds, and improved loop-closure detection, generating high-fidelity three-dimensional maps that were converted into two-dimensional occupancy grids for A-Star global path planning and Dynamic Window Approach (DWA) local control. In the cultivation area, where rails intersect with internal corridors, YOLOv8n-based rail-center detection combined with a pure-pursuit controller established a vision-servo framework for lateral rail switching and inter-row navigation. Field experiments demonstrated that the optimized mapping reduced the mean relative error by 15%. At a navigation speed of 0.2 m/s, the robot achieved a mean lateral deviation of 4.12 cm and a heading offset of 1.79°, while the vision-servo rail-switching system improved efficiency by 25.2%. These findings confirm the proposed framework’s accuracy, robustness, and practical applicability, providing strong support for intelligent facility-agriculture operations. Full article

The increasing complexity of urban traffic networks has highlighted the potential of Multi-Agent Reinforcement Learning (MARL) for Traffic Signal Control (TSC). However, most existing MARL methods assume homogeneous observation and action spaces among agents, ignoring the inherent heterogeneity of real-world intersections in topology and signal phasing, which limits their practical applicability. To address this gap, we propose HAPS-PPO (Heterogeneity-Aware Policy Sharing Proximal Policy Optimization), a novel MARL framework for coordinated signal control in heterogeneous road networks. HAPS-PPO integrates two key mechanisms: an Observation Padding Wrapper (OPW) that standardizes varying observation dimensions, and a Dynamic Multi-Strategy Grouping Learning (DMSGL) mechanism that trains dedicated policy heads for agent groups with distinct action spaces, enabling adequate knowledge sharing while maintaining structural correctness. Comprehensive experiments in a high-fidelity simulation environment based on a real-world road network demonstrate that HAPS-PPO significantly outperforms Fixed-time control and mainstream MARL baselines (e.g., MADQN, FMA2C), reducing average delay time by up to 44.74% and average waiting time by 59.60%. This work provides a scalable and plug-and-play solution for deploying MARL in realistic, heterogeneous traffic networks. Full article

Knee osteoarthritis (OA) is a prevalent condition in older adults that often results in impaired gait and balance, increased risk of falls, and reduced quality of life. Conventional clinical assessments may not adequately capture these deficiencies. This study investigated the gait and balance of elderly individuals with knee OA using wearable inertial measurement units (IMUs). Forty-four participants with Kellgren–Lawrence grade 2–3 knee OA (71.23 ± 5.75 years) and forty-five age-matched controls (70.87 ± 4.30 years) completed dynamic balance (balance board), static balance (single-leg stance), ‘timed up and go’ (TUG), and normal walking tasks. Between 2 and 8 IMUs, depending on the task, were placed on the head, chest, waist, knees, ankles, soles, and balance board to record kinematic data. Balance was quantified using absolute angular velocity and linear acceleration, with group differences analyzed by MANOVA and Bonferroni-adjusted univariate tests. The participants with knee OA exhibited greater gait asymmetry, although the difference was not significant. However, they consistently demonstrated higher absolute angular velocities than controls across most body segments during static and dynamic tasks, indicating reduced postural stability. No group differences were observed in TUG performance. These findings suggest that IMU-based measures, particularly angular velocity, are sensitive to balance impairment detection in knee OA. Incorporating IMU technology into clinical assessments may facilitate early identification of instability and guide targeted interventions to reduce fall risk. Full article

Hearing-impaired people face challenges in expressing and perceiving emotions, and traditional single-modal emotion recognition methods demonstrate limited effectiveness in complex environments. To enhance recognition performance, this paper proposes a multimodal fusion neural network based on a multimodal multi-head attention fusion neural network (MMHA-FNN). This method utilizes differential entropy (DE) and bilinear interpolation features as inputs, learning the spatial–temporal characteristics of brain regions through an MBConv-based module. By incorporating the Transformer-based multi-head self-attention mechanism, we dynamically model the dependencies between EEG and facial expression features, enabling adaptive weighting and deep interaction of cross-modal characteristics. The experiment conducted a four-classification task on the MED-HI dataset (15 subjects, 300 trials). The taxonomy included happy, sad, fear, and calmness, where ‘calmness’ corresponds to a low-arousal neutral state as defined in the MED-HI protocol. Results indicate that the proposed method achieved an average accuracy of 81.14%, significantly outperforming feature concatenation (71.02%) and decision layer fusion (69.45%). This study demonstrates the complementary nature of EEG and facial expressions in emotion recognition among hearing-impaired individuals and validates the effectiveness of feature layer interaction fusion based on attention mechanisms in enhancing emotion recognition performance. Full article

To address the challenges of low detection accuracy and weak generalization in UAV aerial imagery caused by complex ground environments, significant scale variations among targets, dense small objects, and background interference, this paper proposes an improved lightweight multi-scale small-object detection model, MBD-YOLO (MBFF module, BiMS-FPN, and Dual-Stream Head). Specifically, to enhance multi-scale feature extraction capabilities, we introduce the Multi-Branch Feature Fusion (MBFF) module, which dynamically adjusts receptive fields through parallel branches and adaptive depthwise convolutions, expanding the receptive field while preserving detail perception. We further design a lightweight Bidirectional Multi-Scale Feature Aggregation Pyramid Network (BiMS-FPN), integrating bidirectional propagation paths and a Multi-Scale Feature Aggregation (MSFA) module to mitigate feature spatial misalignment and improve small-target detection. Additionally, the Dual-Stream Head with NMS-free architecture leverages a task-aligned architecture and dynamic matching strategies to boost inference speed without compromising accuracy. Experiments on the VisDrone2019 dataset demonstrate that MBD-YOLO-n surpasses YOLOv8n by 6.3% in mAP50 and 8.2% in mAP50–95, with accuracy gains of 17.96–55.56% for several small-target categories, while increasing parameters by merely 3.1%. Moreover, MBD-YOLO-s achieves superior detection accuracy, efficiency, and generalization with only 12.1 million parameters, outperforming state-of-the-art models and proving suitable for resource-constrained embedded deployment scenarios. The superior performance of MBD-YOLO, which harmonizes high precision with low computational demand, fulfills the critical requirements for real-time deployment on resource-limited UAVs, showing great promise for applications in traffic monitoring, urban security, and agricultural surveying. Full article

Virtual commands are significant to model human–computer interactions in autopilot flight missions. However, the huge system hysteresis makes it difficult for proportional–integral–derivative (PID) algorithms to generate the commands that promise better flight convergence. An adaptive trigger compensation neural network method is proposed to dynamically tune the PID parameters, simulating the process of deciding virtual heading commands and performing heading adjustments for virtual pilots. The method consists of trigger filtering, dynamic updating, and compensation synthesis. First, the necessary historical errors are adaptively selected by the threshold trigger filter for better error utilization. Second, error-based initialization is introduced in the neural network PID update process to improve adaptiveness in the initial settings of PID parameters. Third, the parameters are synthesized via error compensation to compute virtual heading commands for acquiring more convergent flight trajectories. The adaptive filter, error-based initialization, and compensation are important to improve the backward propagation neural network in tuning PID parameters. The results demonstrate the advance of the method in simulating heading adjustment behaviors and reducing flight trajectory deviation and fluctuation. The adaptive trigger compensation neural network can enhance the convergent performance of the PID algorithm during autopilot flight scenarios. Full article