Search Results (81)

Wildfire events are increasing in frequency and severity, creating an urgent need for early, accurate, and energy-efficient forest fire prediction systems that can operate at a large scale. A fundamental challenge in edge-level forest fire prediction lies in jointly achieving high detection accuracy while minimizing wireless transmissions and communication-related energy consumption. This paper proposes a communication-aware hierarchical wireless sensor network (WSN) framework that performs fire versus normal environmental state classification directly at the network edge. Multi-modal physical and constrained virtual sensor readings are fused into short-term temporal supervectors and processed locally using lightweight random forest classifiers deployed on sensor nodes and cluster heads. A temporal 2-of-3 voting mechanism is applied at the edge to suppress transient noise and improve prediction reliability before triggering communication. The proposed design enables selective, event-driven transmission, where only temporally validated abnormal states are forwarded through the hierarchy, thereby decoupling detection accuracy from continuous data reporting. Extensive experiments using real multi-modal environmental sensor data and statistically rigorous 5-fold GroupKFold cross-validation—ensuring strict node-level separation between training and testing—demonstrate the effectiveness of the approach. The proposed framework achieves a node-level accuracy of 98.82 ± 1.75% and a scenario-level detection accuracy of 96.52 ± 0.89%. Compared to periodic reporting and the LEACH protocol, the system reduces wireless transmissions by over 66% and communication-related energy consumption by more than 66% across network sizes ranging from 100 to 1000 nodes. The main contributions of this work are summarized as follows: (1) a communication-aware hierarchical Edge-AI framework for early forest fire prediction that performs local inference and temporal validation directly at sensor nodes; (2) a constrained virtual sensing strategy integrated with temporal supervector modeling to enhance spatial coverage while preserving reliability; and (3) a statistically rigorous large-scale evaluation demonstrating joint optimization of prediction accuracy, transmission reduction, and communication energy efficiency across network sizes ranging from 100 to 1000 nodes. These results show that accurate early forest fire prediction can be achieved through edge-level inference and selective communication, substantially extending network lifetime while maintaining statistically reliable detection performance. Full article

Introduction: Rapid and reliable neurovascular imaging is critical for time-sensitive diagnosis in acute cerebrovascular disorders, yet conventional magnetic resonance imaging (MRI) workflows remain constrained by acquisition speed, motion sensitivity, and limited integration of physiological context. We introduce the Intelligent Neurovascular Imaging Engine (INIE), a sensor-informed, topology-aware framework that jointly optimizes accelerated data acquisition, physics-grounded reconstruction, and cross-scale physiological consistency. Methods: INIE combines adaptive sampling, structured low-rank (Hankel) priors, and topology-preserving objectives with multimodal physiological sensors and scanner telemetry, enabling phase-consistent gating and confidence-weighted reconstruction under realistic operating conditions. The framework was evaluated using synthetic phantoms, a translational porcine stroke recanalization model with repeated measures, and retrospective human datasets. Across $N_{\mathrm{runs}}=120$ acquisition–reconstruction runs derived from $N_{\mathrm{animals}}=18$ pigs with animal-level train/validation/test separation, performance was assessed using image quality, topological fidelity, and cross-modal consistency metrics. Multiple-comparison control was performed using Bonferroni/Holm–Bonferroni procedures. Results: INIE achieved acquisition acceleration exceeding 70% while maintaining high reconstruction fidelity (PSNR $\approx 35$–36 dB, SSIM $\approx 0.90$–0.92). Topology-aware analysis showed an approximately twofold reduction in Betti number deviation relative to baseline accelerated methods. Cross-modal validation in a PET subset demonstrated strong agreement between MRI-derived perfusion parameters and metabolic markers (Pearson $r\approx 0.9$). INIE improved large-vessel occlusion detection accuracy to approximately 93% and reduced automated time-to-decision to under three minutes. Conclusions: These results indicate that sensor-informed, topology-aware, closed-loop imaging improves the reliability and physiological consistency of accelerated neurovascular MRI and supports faster, more robust decision-making in acute cerebrovascular imaging workflows. Full article

Natural fractures are critical controls on shale oil storage and migration in the Upper Triassic Chang 7 Member of the Ordos Basin. However, conventional identification techniques—such as mud-invasion correction, R/S rescaled range analysis, and radioactive element analysis—are time-consuming, computationally intensive, and highly dependent on specialized logging data, limiting their large-scale application. To overcome these challenges, this study develops a multi-modal deep neural network that integrates conventional well logs with borehole imaging data. A coupled convolutional neural network (CNN) and deep neural network (DNN) architecture was constructed to predict fracture occurrence, dip angle, and aperture. The model achieves dip-angle prediction accuracies of 98.82% for both training and testing datasets, while aperture prediction accuracies reach 95.97% and 95.91%, respectively. Predicted dip angles are concentrated between 65° and 80°, deviating by less than 0.48° from measured values, whereas apertures fall mainly within 0.5–4.5 cm, with deviations below 0.21 cm except in extreme cases. The CNN branch effectively extracts spatial features from imaging logs, while the DNN branch captures nonlinear relationships in conventional logs. The integrated framework substantially improves fracture characterization accuracy and efficiency. This study provides a scalable and cost-effective approach for rapid fracture identification based on conventional logging data, reducing reliance on specialized imaging logs and supporting integrated geological and engineering evaluations in shale oil reservoirs. Full article

Estimating grape yields in viticulture is hindered by persistent challenges, including strong occlusion between grapes, irregular cluster morphologies, and fluctuating illumination throughout the growing season. This study introduces YOLOv11-IMP, an improved multiscale anchor-free detection framework extending YOLOv11, tailored to vineyard environments. Its architecture comprises five specialized components: (i) a viticulture-oriented backbone employing cross-stage partial fusion with depthwise convolutions for enriched feature extraction, (ii) a bifurcated neck enhanced by large-kernel attention to expand the receptive field coverage, (iii) a scale-adaptive anchor-free detection head for robust multiscale localization, (iv) a cross-modal processing module integrating visual features with auxiliary textual descriptors to enable fine-grained cluster-level yield estimation, and (v) aross multiple scales. This work evaluated YOLOv11-IMP on five grape varieties collecten augmented spatial pyramid pooling module that aggregates contextual information acd under diverse environmental conditions. The framework achieved 94.3% precision and 93.5% recall for cluster detection, with a mean absolute error (MAE) of 0.46 kg per vine. The robustness tests found less than 3.4% variation in accuracy across lighting and weather conditions. These results demonstrate that YOLOv11-IMP can deliver high-fidelity, real-time yield data, supporting decision-making for precision viticulture and sustainable agricultural management. Full article

Distribution shifts commonly arise in real-world machine learning scenarios in which the fundamental assumption that training and test data are drawn from independent and identically distributed samples is violated. In the case of medical data, such distribution shifts often occur during data acquisition and pose a significant challenge to the robustness and reliability of artificial intelligence systems in clinical practice. Additionally, quantifying these shifts without training a model remains a key open problem. This paper proposes a comprehensive methodological framework for evaluating the impact of such shifts on medical image datasets under artificial transformations that simulate acquisition variations, leveraging the Cumulative Spectral Gradient (CSG) score as a measure of multiclass classification complexity induced by distributional changes. Building on prior work, the proposed approach is meaningfully extended to twelve 2D medical imaging benchmarks from the MedMNIST collection, covering both binary and multiclass tasks, as well as grayscale and RGB modalities. We evaluate the metric analyzing its robustness to clinically inspired distribution shifts that are systematically simulated through motion blur, additive noise, brightness and contrast variation, and sharpness variation, each applied at three severity levels. This results in a large-scale benchmark that enables a detailed analysis of how dataset characteristics, transformation types, and distortion severity influence distribution shifts. Thus, the findings show that while the metric remains generally stable under noise and focus distortions, it is highly sensitive to variations in brightness and contrast. On the other hand, the proposed methodology is compared against Cleanlab’s widely used Non-IID score on the RetinaMNIST dataset using a pre-trained ResNet-50 model, including both class-wise analysis and correlation assessment between metrics. Finally, interpretability is incorporated through class activation map analysis on BloodMNIST and its corrupted variants to support and contextualize the quantitative findings. Full article

Recent years have witnessed performance improvements in Multimodal Large Language Models (MLLMs) on downstream natural image understanding tasks. However, when applied to the function graph reasoning task, which is highly information-dense and abundant in fine-grained structural details, these models face pronounced performance degradation. The challenges are primarily characterized by several core issues: the static projection bottleneck, inadequate cross-modal interaction, and insufficient visual context in text embeddings. To address these problems, this study proposes a multimodal feature fusion enhancement method for function graph reasoning and constructs the FuncFusion-Math model. The core innovation of this model resides in its design of a dual-path feature fusion mechanism for both image and text. Specifically, the image fusion module adopts cross-attention and self-attention mechanisms to optimize visual feature representations under the guidance of textual semantics, effectively mitigating fine-grained information loss. The text fusion module, through feature concatenation and Transformer encoding layers, deeply integrates structured mathematical information from the image into the textual embedding space, significantly reducing semantic deviation. Furthermore, this study utilizes a four-stage progressive training strategy and incorporates the LoRA technique for parameter-efficient optimization. Experimental results demonstrate that the FuncFusion-Math model, with 3B parameters, achieves an accuracy of 43.58% on the FunctionQA subset of the MathVista test set, outperforming a 7B-scale baseline model by 13.15%, which validates the feasibility and effectiveness of the proposed method. Full article

This study systematically verified the applicability and accuracy of the Probability Density Evolution Method (PDEM) in the probabilistic modeling of the dynamic response of anchored rock slopes under random seismic action through large-scale shaking table model tests. Across 144 sets of non-stationary random ground motions and 7 sets of white noise excitations, key response data such as acceleration, displacement, and changes in anchor axial force were collected. The PDEM was used to model the instantaneous probability density function (PDF) and cumulative distribution function (CDF), which were then compared with the results of normal distribution, Gumbel distribution, and direct sample statistics from multiple dimensions. The results show that the PDEM does not require a preset distribution form and can accurately reproduce the non-Gaussian, multi-modal, and time evolution characteristics of the response; in the reliability assessment of peak responses, its prediction deviation is much smaller than that of traditional parametric models; the three-dimensional probability density evolution cloud map further reveals the law governing the entire process of the response PDF from “narrow and high” in the early stage of the earthquake, “wide and flat” in the main shock stage, to “re-convergence” after the earthquake. The study confirms that the PDEM has significant advantages and engineering application value in the analysis of random seismic responses and the dynamic reliability assessment of anchored slopes. Full article

Accurate indoor positioning and navigation remain significant challenges, with audio sensor-based sound source localization emerging as a promising sensing modality. Conventional methods, often reliant on multi-channel processing or time-delay estimation techniques such as Generalized Cross-Correlation, encounter difficulties regarding computational complexity, hardware synchronization, and reverberant environments where time difference in arrival cues are masked. While machine learning approaches have shown potential, their performance depends heavily on the discriminative power of input features. This paper proposes a novel feature extraction method named Short-Time Homomorphic Deconvolution, which transforms multi-channel audio signals into a 2D Time × Time-of-Flight representation. Unlike prior 1D methods, this feature effectively captures the temporal evolution and stability of time-of-flight differences between microphone pairs, offering a rich and robust input for deep learning models. We validate this feature using a lightweight Convolutional Neural Network integrated with a dual-stage channel attention mechanism, designed to prioritize reliable spatial cues. The system was trained on a large-scale dataset generated via simulations and rigorously tested using real-world data acquired in an ISO-certified anechoic chamber. Experimental results demonstrate that the proposed model achieves precise Direction of Arrival estimation with a Mean Absolute Error of 1.99 degrees in real-world scenarios. Notably, the system exhibits remarkable consistency between simulation and physical experiments, proving its effectiveness for robust indoor navigation and positioning systems. Full article

Background: Colorectal cancer (CRC) is the second leading cause of cancer-related mortality worldwide, accounting for approximately 10% of all cancer cases. Despite the proven effectiveness of conventional screening modalities such as colonoscopy and fecal immunochemical testing (FIT), their invasive nature, high cost, and limited patient compliance hinder widespread adoption. Recent advancements in artificial intelligence (AI) and bowel sound-based signal processing have enabled non-invasive approaches for gastrointestinal diagnostics. Among these, bowel sound analysis—historically considered subjective—has reemerged as a promising biomarker using digital auscultation and machine learning. Objective: This review explores the potential of AI-powered bowel sound analytics for early detection, screening, and characterization of colorectal cancer. It aims to assess current methodologies, summarize reported performance metrics, and highlight translational opportunities and challenges in clinical implementation. Methods: A narrative review was conducted across PubMed, Scopus, Embase, and Cochrane databases using the terms colorectal cancer, bowel sounds, phonoenterography, artificial intelligence, and non-invasive diagnosis. Eligible studies involving human bowel sound-based recordings, AI-based sound analysis, or machine learning applications in gastrointestinal pathology were reviewed for study design, signal acquisition methods, AI model architecture, and diagnostic accuracy. Results: Across studies using convolutional neural networks (CNNs), gradient boosting, and transformer-based models, reported diagnostic accuracies ranged from 88% to 96%. Area under the curve (AUC) values were ≥0.83, with F1 scores between 0.71 and 0.85 for bowel sound classification. In CRC-specific frameworks such as BowelRCNN, AI models successfully differentiate abnormal bowel sound intervals and spectral patterns associated with tumor-related motility disturbances and partial obstruction. Distinct bowel sound-based signatures—such as prolonged sound-to-sound intervals and high-pitched “tinkling” proximal to lesions—demonstrate the physiological basis for CRC detection through bowel sound-based biomarkers. Conclusions: AI-driven bowel sound analysis represents an emerging, exploratory research direction rather than a validated colorectal cancer screening modality. While early studies demonstrate physiological plausibility and technical feasibility, no large-scale, CRC-specific validation studies currently establish sensitivity, specificity, PPV, or NPV for cancer detection. Accordingly, bowel sound analytics should be viewed as hypothesis-generating and potentially complementary to established screening tools, rather than a near-term alternative to validated modalities such as FIT, multitarget stool DNA testing, or colonoscopy. Full article

Semantic segmentation of workpieces and die cavities is critical for intelligent process monitoring and quality control in hammer die-forging. However, the field of 3D point cloud segmentation currently faces prominent limitations in forging scenario adaptation: existing state-of-the-art (SOTA) methods are predominantly optimized for road driving or indoor scenes, where targets have stable poses and regular surfaces. They lack dedicated designs for capturing the fine-grained deformation characteristics of forging workpieces and alleviating multi-scale feature misalignment caused by large pose variations—key pain points in forging segmentation. Consequently, these methods fail to balance segmentation accuracy and real-time efficiency required for practical forging applications. To address this gap, this paper proposes a novel semantic segmentation framework fusing 3D point cloud and bird’s-eye-view (BEV) representations for complex die-forging scenes. Specifically, a Star-based encoding module is designed in the BEV encoding stage to enhance capture of fine-grained workpiece deformation characteristics. A hierarchical feature-offset alignment mechanism is developed in decoding to alleviate multi-scale spatial and semantic misalignment, facilitating efficient cross-layer fusion. Additionally, a weighted adaptive fusion module enables complementary information interaction between point cloud and BEV modalities to improve precision.We evaluate the proposed method on our self-constructed simulated and real die-forging point cloud datasets. The results show that when trained solely on simulated data and tested directly in real-world scenarios, our method achieves an mIoU that surpasses RPVNet by $1.1\%$. After fine-tuning with a small amount of real data, the mIoU further improves by $5\%$, reaching optimal performance. Full article

This study aims to demonstrate the feasibility of a hybrid measurement system that combines Laser Doppler Vibrometry (LDV) and conventional accelerometers for operational modal analysis (OMA) of civil engineering structures. The proposed approach addresses the limitations of traditional accelerometer-based systems, particularly for large-scale or inaccessible structures, by integrating non-contact LDV measurements with conventional sensor data. Experimental tests were conducted on a cantilever beam and a pedestrian laboratory footbridge to validate the hybrid system. The LDV was used to measure velocity at key points, while accelerometers provided complementary reference acceleration measurements. Reflective targets were employed to facilitate non-contact data collection, allowing for the subsequent reuse of these targets for repeated measurements. The velocity data from the LDV were differentiated to obtain acceleration and integrated to estimate displacement, enabling a direct combination with accelerometer data. ARTeMIS Modal software was utilized to process and analyze the collected data, successfully identifying the natural frequencies and vibration modes of both structures. The results demonstrate that the LDV–accelerometer hybrid system effectively captures the dynamic behavior of structures, offering a comprehensive solution for modal analysis without extensive sensor deployment. This approach provides significant advantages in scenarios where traditional methods are impractical, positioning the hybrid system as a promising tool for dynamic analysis and infrastructure monitoring of complex structures. Full article

Automated animal identification is a practical task for reuniting lost pets with their owners, yet current systems often struggle due to limited dataset scale and reliance on unimodal visual cues. This study introduces a multimodal verification framework that enhances visual features with semantic identity priors derived from synthetic textual descriptions. We constructed a massive training corpus of 1.9 million photographs covering 695,091 unique animals to support this investigation. Through systematic ablation studies, we identified SigLIP2-Giant and E5-Small-v2 as the optimal vision and text backbones. We further evaluated fusion strategies ranging from simple concatenation to adaptive gating to determine the best method for integrating these modalities. Our proposed approach utilizes a gated fusion mechanism and achieved a Top-1 accuracy of 84.28% and an Equal Error Rate of 0.0422 on a comprehensive test protocol. These results represent an 11% improvement over leading unimodal baselines and demonstrate that integrating synthesized semantic descriptions significantly refines decision boundaries in large-scale pet re-identification. Full article

Addressing the issue of fruit recognition and localization failures in harvesting robots due to severe occlusion by branches and leaves in complex orchard environments, this paper proposes an occlusion avoidance method that combines a lightweight YOLOv8n model, developed by Ultralytics in the United States, with active perception. Firstly, to meet the stringent real-time requirements of the active perception system, a lightweight YOLOv8n model was developed. This model reduces computational redundancy by incorporating the C2f-FasterBlock module and enhances key feature representation by integrating the SE attention mechanism, significantly improving inference speed while maintaining high detection accuracy. Secondly, an end-to-end active perception model based on ResNet50 and multi-modal fusion was designed. This model can intelligently predict the optimal movement direction for the robotic arm based on the current observation image, actively avoiding occlusions to obtain a more complete field of view. The model was trained using a matrix dataset constructed through the robot’s dynamic exploration in real-world scenarios, achieving a direct mapping from visual perception to motion planning. Experimental results demonstrate that the proposed lightweight YOLOv8n model achieves a mAP of 0.885 in apple detection tasks, a frame rate of 83 FPS, a parameter count reduced to 1,983,068, and a model weight file size reduced to 4.3 MB, significantly outperforming the baseline model. In active perception experiments, the proposed method effectively guided the robotic arm to quickly find observation positions with minimal occlusion, substantially improving the success rate of target recognition and the overall operational efficiency of the system. The current research outcomes provide preliminary technical validation and a feasible exploratory pathway for developing agricultural harvesting robot systems suitable for real-world complex environments. It should be noted that the validation of this study was primarily conducted in controlled environments. Subsequent work still requires large-scale testing in diverse real-world orchard scenarios, as well as further system optimization and performance evaluation in more realistic application settings, which include natural lighting variations, complex weather conditions, and actual occlusion patterns. Full article

Acoustic metamaterials are artificially engineered materials composed of subwavelength structural units, whose effective acoustic properties are primarily determined by structural design rather than intrinsic material composition. By introducing local resonances, these materials can exhibit unconventional acoustic behavior, enabling enhanced sound insulation beyond the limitations of conventional structures. In this study, a thin plate (thin sheet) refers to a structural element whose thickness is much smaller than its in-plane dimensions and can be accurately described using classical thin-plate vibration theory. When resonant mass blocks are attached to a thin plate, a thin-plate acoustic metamaterial is formed through the coupling between plate bending vibrations and local resonances. Thin-plate acoustic metamaterials exhibit excellent sound insulation performance in the low- and mid-frequency ranges. Multilayer configurations and the combination with porous materials can effectively broaden the insulation bandwidth and improve overall performance. However, the large number of structural parameters in multilayer composite thin-plate acoustic metamaterials significantly increases design complexity, making conventional trial-and-error approaches inefficient. To address this challenge, a neural-network-based inverse design framework is proposed for multilayer composite thin-plate acoustic metamaterials. An analytical model of thin-plate metamaterials with multiple attached cylindrical masses is established using the point matching and modal superposition methods and validated by finite element simulations. A multilayer composite unit cell is then constructed, and a dataset of 30,000 samples is generated through numerical simulations. Based on this dataset, a forward prediction network achieves a test error of 1.06%, while the inverse design network converges to an error of 2.27%. The inverse-designed structure is finally validated through impedance tube experiments. The objective of this study is to establish a systematic theoretical and neural-network-assisted inverse design framework for multilayer thin-plate acoustic metamaterials. The main novelties include the development of an accurate analytical model for thin-plate metamaterials with multiple attached masses, the construction of a large-scale simulation dataset, and the proposal of a neural-network-assisted inverse design strategy to address non-uniqueness in inverse design. The proposed approach provides an efficient and practical solution for low-frequency sound insulation design. Full article

Background and Objectives: Nasopharyngeal carcinoma (NPC) displays marked geographic and histopathological heterogeneity, and prognostic determinants in non-endemic regions remain incompletely defined. This study aimed to evaluate the impact of clinicopathological characteristics and treatment modalities on survival outcomes among patients with stage II–IVA NPC treated with curative intent at a single tertiary cancer center. Materials and Methods: A retrospective analysis was conducted on 81 consecutive patients with histologically confirmed NPC treated between 2000 and 2022. Demographic, clinical, and treatment parameters were extracted from institutional records. Survival outcomes—including disease-free survival (DFS), locoregional recurrence-free survival (LRFS), distant metastasis-free survival (DMFS), cancer-specific survival (CSS), and overall survival (OS)—were estimated using the Kaplan–Meier method and compared using the log-rank test. Prognostic variables identified in univariate analysis were further assessed by multivariable Cox proportional hazards regression (Cox’s model). Results: The cohort included 59 men (72.8%) and 22 women (27.2%), with a median age of 50.8 years (range, 19–78). Most patients presented with locally advanced disease (T3–T4, 53.1%; N2, 60.5%; stage III–IVA, 87.7%). Non-keratinizing undifferentiated carcinoma (World Health Organization [WHO] type III) was the predominant histology (71.6%), followed by the non-keratinizing differentiated subtype (17.3%). Median DFS and OS were 94.6 and 139.4 months, respectively. According to the univariate analysis, histological subtypes and a family history of cancer were significantly associated with DFS, whereas comorbid systemic disease showed an unexpected association with longer DMFS. The multivariable Cox model identified the histological subtype as an independent predictor of disease recurrence (HR = 2.23, 95% CI: 1.00–4.94; p = 0.049). For OS, both histological subtype (HR = 2.40, 95% CI: 1.10–5.25; p = 0.029) and age at diagnosis (HR = 1.05, 95% CI: 1.02–1.09; p = 0.005) were independent adverse prognostic factors. Conclusions: In this long-term, single-center study from a non-endemic region, histological subtype emerged as the most powerful determinant of prognosis, significantly influencing both DFS and OS. Patients with non-keratinizing undifferentiated (WHO type III) carcinoma demonstrated superior outcomes compared with those with differentiated histology. Additionally, increasing age at diagnosis was independently associated with poorer OS. In contrast, inflammatory and nutritional biomarkers, the Pan-Immune–Inflammation Value (PIV) and the Prognostic Nutritional Index (PNI), showed no prognostic significance. These findings underscore the continued prognostic relevance of histopathologic classification and age and highlight the need for large-scale, standardized studies integrating Epstein–Barr virus (EBV) status and host-related factors in non-endemic NPC populations. Full article