Search Results (16,952)

Free space detection in assisted driving applications is essential to provide information to vehicles about traversable surfaces and potential obstacles to be avoided. The current trend in free space detection favors the use of deep learning techniques. However, Deep Neural Networks require extensive training that considers as many scenarios as possible, which makes it difficult to create a model that can be generalized to all types of surfaces. Additionally, their lack of explainability contrasts with the growing interest in geometrically grounded and safety-oriented design principles for autonomous vehicle systems. To address these limitations, we propose a geometric approach that incorporates coplanarity conditions and normal vector estimation, removing the dependence on datasets for different types of surfaces. Additionally, the stereoscopic images are clustered in superpixels. The use of images clustered in superpixels allows us to obtain shorter processing times, in addition to taking advantage of the spatial and color information provided by the superpixels to increase the robustness of the three-dimensional reconstruction of the scene. Experimental results show that the proposed superpixel-based approach achieves competitive performance compared to unsegmented dense stereo methods, while significantly reducing algorithmic complexity. These results demonstrate the viability of integrating superpixel clustering into stereo-based free space estimation frameworks. Full article

Remaining Useful Life (RUL) prediction is a key component of prognostics and health management for modern industrial systems. While deep learning methods have significantly improved prediction accuracy, many existing approaches rely on large neural networks that are difficult to deploy on resource-constrained edge devices. This study presents a deployment-oriented evaluation of compact neural networks for RUL prediction using the NASA C-MAPSS turbofan engine benchmark. Two lightweight hybrid architectures, CNN–GRU and CNN–TCN, were developed with approximately 28k–32k parameters to represent realistic models for CPU-based edge inference. A systematic experimental analysis was conducted across all four C-MAPSS subsets (FD001–FD004), which represent increasing levels of operational and fault complexity. In addition to baseline performance, two post-training compression techniques (i.e., global unstructured magnitude pruning and dynamic INT8 quantization) were evaluated. To assess real deployment behavior, inference latency was measured on both a high-performance Intel x86 workstation and a resource-constrained ARM platform. Results show that CNN–GRU generally achieves higher predictive accuracy, whereas CNN–TCN provides more consistent and lower inference latency due to its convolution-only temporal modeling. Unstructured pruning can yield modest improvements in prediction accuracy, suggesting a regularization effect, but it does not reliably reduce model size or latency on standard CPUs due to the overhead associated with pruning masks. Dynamic quantization substantially reduces model size (particularly for CNN–GRU) while preserving predictive accuracy; however, it increases runtime latency because of additional quantization and dequantization operations. These findings demonstrate that compression techniques commonly used for large models do not necessarily translate into deployment benefits for already compact RUL architectures and highlight the importance of hardware-aware evaluation when designing edge prognostics systems. Full article

Low-count positron emission tomography (PET) is inherently affected by Poisson-dominated noise, which degrades image contrast, structural delineation, and quantitative reliability. This study systematically evaluated residual learning-based deep neural networks to investigate the influence of residual block depth on PET image restoration performance under low-count conditions. We employed a physically controlled striatum phantom, fabricated using 3D printing technology, to ensure reproducible acquisition conditions and controlled physical variability. PET images were acquired using a clinical PET/computed tomography (CT) system with list-mode acquisition. Low-count images reconstructed from short-duration acquisition were paired with high-count reference images reconstructed from extended acquisitions. We compared conventional filtering techniques, including median, Wiener, and modified median Wiener filters, with residual network (ResNet)-based models incorporating 8, 16, and 32 residual blocks. Image quality was quantitatively assessed using contrast-to-noise ratio (CNR), coefficient of variation (COV), line profile analysis, universal quality index (UQI), and perceptual image patch similarity (LPIPS). The results demonstrated that ResNet-based restorations substantially outperformed conventional filtering techniques in contrast recovery, signal stability, and structural preservation. The ResNet-16 model achieved the most balanced performance, yielding the highest CNR (9.02) and lowest COV (0.105), while also demonstrating superior structural and perceptual similarity, as indicated by UQI (0.9224) and LPIPS (0.0174), relative to the high-count reference images. Deeper network configurations exhibited diminishing returns and reduced structural consistencies. These findings indicate that an intermediate residual block depth is optimal for low-count PET image restoration and highlight the importance of architectural optimization in deep learning-based PET image enhancement with phantom-based evaluation frameworks. Full article

The increasing proportion of renewable energy leads to frequent changes in turbine load, making the regulating stage more prone to degradation. Traditional anomaly detection methods lack sufficient sensitivity and generalization. To address this issue, this study proposes a method combining multi-attention gated fusion and physical information learning. A gated fusion mechanism is proposed to adaptively extract and fuse key temporal and feature information. Furthermore, the generalization ability of the model is improved by introducing physical constraints derived from the relationship between pressure, temperature, and valve position. Finally, a dynamic temperature prediction model is established using the multi-output long short-term memory neural network. Experiments using actual power plant data demonstrate that the proposed method effectively improves the accuracy of post-regulating-stage temperature prediction and the sensitivity of anomaly detection. The proposed gating fusion method improves prediction accuracy by 4.6% compared to direct addition, while the fusion of physical information reduces the generalization error by more than 6%. In addition, compared to traditional deep learning and machine learning models, the proposed method improves anomaly detection accuracy by at least 3.9%. This research is of great significance for the safe operation of thermal power units and the power grid. Full article

Computed tomography (CT) enables non-destructive insight into internal log structure, yet fully automated interpretation of CT images remains limited by inconsistent annotations, boundary ambiguity, and insufficient spatial context in 2D slice-based analysis. These challenges restrict the industrial deployment of deep learning for wood quality assessment. This study applies artificial intelligence (AI) and deep learning to the automated analysis of computed tomography (CT) scans of wood logs for detecting internal qualitative features and segmenting bark. Using convolutional neural networks (CNNs), trained models accurately distinguish healthy and damaged regions and segment bark, including discontinuous parts. We introduce a novel pseudo-spatial representation by merging consecutive slices into red–green–blue (RGB) format, which improves prediction accuracy and model robustness across logs. To enhance interpretability, Gradient-weighted Class Activation Mapping (Grad-CAM) highlights regions contributing most to defect detection, particularly knots. Comprehensive evaluation using Sørensen–Dice similarity coefficients and confusion matrices confirms the effectiveness of the proposed approach under industrial conditions. These findings demonstrate that AI-driven CT image analysis can address key limitations of current log-grading workflows and enable more reliable, objective, and scalable quality assessment for timber-dependent economies. Full article

Early detection of battery degradation is essential for ensuring the safety and reliability of electric vehicle (EV) systems under real-world operating variability. This paper proposes a physics-guided multi-sensor learning framework, termed SensorFusion-Former (SFF), for early warning of short-term EV battery performance degradation. The proposed approach integrates a physics-based baseline model for operational normalization, a multi-sensor fusion attention mechanism to model cross-modality interactions, and a lightweight transformer architecture for efficient temporal representation learning. Weak supervision is derived from physics-consistent residual analysis with temporal smoothing, enabling scalable training without dense manual annotations. To support reliable deployment, evidential uncertainty modeling and conformal calibration are incorporated to obtain statistically controlled decision thresholds. Experiments conducted on a real driving cycle dataset from IEEE DataPort demonstrate that SFF consistently outperforms classical machine learning methods, deep neural networks, and standard transformer models in terms of early-warning lead time, false alarm rate, and inference efficiency while maintaining competitive discriminative performance. Cross-scenario evaluations under diverse thermal conditions further confirm the robustness and generalization capability of the proposed framework. Full article

The transition to smart cities is accelerating distributed wind and solar deployment. However, their intermittency challenges grid operation, thereby making accurate machine-learning-based prediction of wind speed and global horizontal irradiance (GHI) crucial. This study presents a cost-effective approach that enhances prediction accuracy by extracting additional features from timestamp records for deep learning models used to forecast GHI and wind speed. Unlike conventional methods that require onsite meteorological measurements, the proposed approach uses only date and time information as inputs to multivariate deep neural networks, including recurrent neural networks, gated recurrent units, long short-term memory (LSTM), bidirectional LSTM, and convolutional neural networks. For wind speed prediction, the proposed configuration achieves R² up to 0.9987, with RMSE as low as 0.067 m/s for 3 d ahead forecasting, outperforming univariate baselines and matching models. For GHI forecasting, the time-based configuration attains R² values above 0.9994 in 12 h ahead predictions, with the RMSE reduced to approximately 4.47 W/m², representing a substantial improvement over univariate models. The proposed framework maintains strong performance, particularly under clear and sunny conditions. These results demonstrate that timestamp-engineered features can deliver forecasting accuracy comparable to conventional multivariate meteorological models while significantly reducing infrastructure requirements, making the approach well-suited for scalable smart city energy management. Full article

Vertical docking of the autonomous underwater shuttle (AUS) for deep-sea data relay relies heavily on ultra-short baseline (USBL) acoustic positioning, whose measurements can be intermittently unavailable and contaminated by outliers in complex underwater environments. This paper proposes a neural network-enhanced robust navigation framework to improve AUS navigation reliability during acoustically guided vertical docking under USBL outages. First, a model-aided batch maximum a posteriori trajectory estimation method (MA-BMAP) is developed to generate learning quality supervision under sensor-limited conditions. Based on the estimated trajectories, a long short-term memory (LSTM)-based horizontal velocity predictor is integrated into a robust fusion filter with online ocean current estimation, enabling stable state estimation during USBL outages and robust rejection of abnormal USBL measurements. The proposed framework is validated through simulations and field trials in lake and sea environments. In sea trials, during two representative 200 s USBL outage intervals, the end-of-window horizontal position errors are 7.86 m and 4.14 m, respectively, corresponding to AUS-to-docking station distances of 244 m and 51 m. In addition, the introduced USBL outliers are successfully detected and rejected. The results indicate that the proposed method enables accurate and stable navigation during USBL unavailability and rapid recovery once USBL measurements resume, demonstrating its practicality for vertical docking missions. Full article

Frequency-domain electromagnetic (FDEM) measurements for shallow subsurface exploration are frequently corrupted by noise, which masks weak secondary-field responses and degrades interpretation. We propose an electromagnetic diffusion CNN (EMDiC) for 1D multi-frequency FDEM denoising, where denoising is formulated as conditional diffusion-based generation. EMDiC combines an analytic frequency–spatial encoder, a Feature-wise Linear Modulation (FiLM)-conditioned convolutional hourglass backbone, and a physics-informed composite loss built on velocity loss to improve waveform reconstruction under severe noise. A reproducible synthetic dataset is constructed through layered-earth forward modeling with concentric Transmitter–Receiver (TX–RX) geometry, multiple target categories, and mixed noise waveforms. On synthetic benchmarks covering multiple noise levels and material types, EMDiC achieves the best overall performance in Root Mean Square Error (RMSE), Signal-to-Noise Ratio (SNR), and Normalized cross-correlation (NCC) among 1D U-Net, diffusion-based variants, and representative neural baselines, with the clearest gains under medium-to-strong noise and for targets with pronounced induction responses. Ablation experiments verify the complementary contributions of electromagnetic positional encoding (EMPE), FiLM conditioning, and the composite loss. Field data validation with a self-developed GEM-3 system further shows that EMDiC improves cross-frequency coherence and suppresses oscillations while preserving the main response characteristics. Full article

This paper aims to present a non-destructive, optimized variational mode decomposition (VMD)-based ground-penetrating radar (GPR) method developed for identifying void defects in reinforced concrete (RC) structures. This study also presents an enhanced framework for defect detection in RC by integrating advanced spectrum analysis with deep learning techniques. A GPR investigation was conducted on an RC bridge deck with known structural defects to generate a representative dataset reflecting both intact and void-defective conditions. In addition to conventional spectral techniques such as fast Fourier transform (FFT), spectrogram, and scalogram, an optimized variational mode decomposition (VMD) method was implemented. The VMD approach decomposes GPR signals into intrinsic mode functions, enabling refined feature extraction beyond traditional spectral methods and allowing clear differentiation between intact and defective signals. The limited availability and quality of GPR small datasets have restricted the application of a functional 1D-CNN which generally requires at least several hundred datasets. To address this challenge, a data augmentation strategy is adopted. FFT-based features were successfully utilized to train a one-dimensional convolutional neural network (1D-CNN) for automated defect identification. The results demonstrate that both the advanced spectrum-based approach and the hybrid framework combining spectral analysis with deep learning significantly improve defect detection performance. Overall, the proposed methodology provides an effective and intelligent solution to support timely, data-driven decision-making for maintenance and safety assurance of bridge infrastructure. Full article

Visual speech recognition systems encounter significant challenges in multilingual nations such as the Philippines, where numerous regional languages, including Cebuano and Ilocano, feature distinct phonetic-visual characteristics. Deep learning models such as the Lip Reading Network and the Lightweight Crowd Segmentation Network have demonstrated strong performance with 3D Convolutional Neural Networks (CNNs). However, their substantial computational requirements restrict deployment on portable edge devices. We introduce a more efficient alternative that integrates a 2D Residual Network 50 architecture with a Long Short-Term Memory network and Connectionist Temporal Classification for lip-reading classification of Philippine regional languages. The proposed model is deployed on the Jetson Orin Nano, a high-performance edge device optimized for real-time inference through Compute Unified Device Architecture acceleration. Using a dataset of 2000 annotated videos encompassing 10 lexicons each for Cebuano and Ilocano, the model’s effectiveness was evaluated. Results achieved a regional language classification accuracy of 90%, with lexicon-level accuracies of 74% for Cebuano and 66% for Ilocano. This work represents a step toward developing accessible and scalable communication aids for deaf communities in linguistically diverse environments, leveraging transfer learning on pretrained models. Full article

Unmanned Aerial Vehicle (UAV) swarms operating in contested environments face a critical “semantic gap” between raw, high-velocity network traffic and high-level mission security constraints, compounded by the risk of privacy leakage during collaborative learning. Existing deep learning (DL)-based Network Intrusion Detection Systems (NIDSs) suffer from opacity, prohibitive resource consumption, and vulnerability to gradient leakage attacks in federated settings, while traditional rule-based systems fail to handle encrypted payloads and evolving attack patterns. To bridge this gap, we present NeSySwarm-IDS (Neuro-Symbolic Swarm Intrusion Detection System), an end-to-end differentiable neuro-symbolic framework that simultaneously achieves high accuracy, strong privacy guarantees, and built-in interpretability under resource constraints. NeSySwarm-IDS integrates an extremely lightweight 1D convolutional neural network with a differentiable Łukasiewicz fuzzy logic reasoner incorporating attack-specific rules. By aggregating only low-dimensional logic rule weights with calibrated differential privacy noise, we drastically reduce communication overhead while providing $(ϵ,\delta )$-DP guarantees with negligible utility loss. Extensive experiments on the UAV-NIDD dataset and our self-collected dataset demonstrate that NeSySwarm-IDS achieves near-perfect detection accuracy, significantly outperforming traditional machine learning baselines despite using limited training data. A detailed case study on GPS spoofing confirms the interpretability of our approach, providing axiomatic explanations suitable for autonomous mission verification. These results establish that end-to-end neuro-symbolic learning can effectively bridge the semantic gap in UAV swarm security while ensuring privacy and interpretability, offering a practical pathway for deploying trustworthy AI in contested environments. Full article

Background/Objectives: Accurate detection of thyroid cancer using ultrasound remains a challenge, as malignant nodules can be microscopic and heterogeneous, easily confused with point clusters and borderline-featured tissues. Current studies in deep learning demonstrate good performance with convolutional neural networks (CNNs) and clustering; however, many approaches focus on local tissue and provide limited, non-quantitative interpretation, reducing clinical confidence. This study proposes an integrated framework combining enhanced convolutional feature encoders (DenseNet169 and VGG19) with an enhanced vision transformer (ViT-E) to integrate local feature and global relational context during learning, rather than delayed integration. Methods: The proposed framework integrates enhanced convolutional feature encoders (DenseNet169 and VGG19) with an enhanced vision transformer (ViT-E), enabling simultaneous learning of local feature representations and global relational context. This design allows feature fusion during the learning stage instead of delayed integration, aiming to improve diagnostic performance and interpretability in thyroid ultrasound image analysis. Results: The best-performing model, ViT-E–DenseNet169, achieved 98.5% accuracy, 98.9% sensitivity, 99.15% specificity, and 97.35% AUC, surpassing the robust basic hybrid model (CNN–XGBoost/ANN) and existing systems. A second contribution is improved interpretability, moving from mere illustration to validation. Gradient-weighted class activation mapping (Grad-CAM) maps demonstrated distinct and clinically understandable concentration patterns across various thyroid cancers: precise intralesional concentration for high-confidence malignancies (PTC = 0.968), edge/interface concentration for capsule risk patterns (PTC = 0.957), and broader-field activation consistent with infiltration concerns (PTC = 0.984), while benign scans showed low and diffuse activation (PTC = 0.002). Spatial audits reinforced this behavior (IoU/PAP: 0.72/91%, 0.65/78%, 0.58/62%). Conclusions: The integrated ViT-E–DenseNet169 framework provides highly accurate thyroid cancer detection while offering clinically meaningful interpretability through Grad-CAM-based spatial validation, supporting improved confidence in AI-assisted ultrasound diagnosis. Full article

Existing autonomous navigation systems for Unmanned Aerial Vehicles (UAVs) face the dual challenges of local minima entrapment and computational complexity that scales with environmental density. This paper proposes a hierarchical navigation architecture integrating deep representation learning with an improved Vortex Artificial Potential Field (APF). At the decision layer, a Convolutional Neural Network (CNN) encodes the environment as a fixed-dimensional tensor and generates global waypoints with constant-time inference, independent of obstacle count. At the control layer, a Vortex APF resolves the Goal Non-Reachable with Obstacles Nearby (GNRON) problem and limit-cycle oscillations through tangential rotational potentials, achieving significant improvement in trajectory smoothness compared to traditional APF methods. A closed-loop replanning mechanism further ensures robust performance under execution drift. Experiments across varying obstacle densities demonstrate that the combined system achieves high navigation success rates in dense environments with substantially reduced computation time compared to sampling-based planners such as Rapidly exploring Random Tree star (RRT*), while maintaining superior trajectory quality. This architecture provides a computationally efficient solution for resource-constrained UAV platforms operating in GPS-denied or obstacle-rich environments such as warehouses, forests, and disaster sites. Full article