Search Results (419)

Multispectral object detection is a fundamental task with an extensive range of practical implications. In particular, combining visible (RGB) and infrared (IR) images can offer complementary information that enhances detection performance in different weather scenarios. However, the existing methods generally involve aligning features across modalities and require proposals for the two-stage detectors, which are often slow and unsuitable for large-scale applications. To overcome this challenge, we introduce a novel one-stage oriented detector for RGB-infrared object detection called the Layer-wise Cross-Modality calibration and Aggregation (LCMA) detector. LCMA employs a layer-wise strategy to achieve cross-modality alignment by using the proposed inter-modality spatial-reduction attention. Moreover, we design Gated Coupled Filter in each layer to capture semantically meaningful features while ensuring that well-aligned and foreground object information is obtained before forwarding them to the detection head. This relieves the need for a region proposal step for the alignment, enabling direct category and bounding box predictions in a unified one-stage oriented detector. Extensive experiments on two challenging datasets demonstrate that the proposed LCMA outperforms state-of-the-art methods in terms of both accuracy and computational efficiency, which implies the efficacy of our approach in exploiting multi-modality information for robust and efficient multispectral object detection. Full article

Problem Statement: Upper crossed syndrome (UCS), as first described by Janda, refers to a group of muscle imbalances in which tightness in the upper trapezius and levator scapulae dorsally cross with tightness in the pectoralis major and minor muscles, and weakness of deep cervical flexors cross ventrally with weakness of the middle and lower trapezius. Postural alterations from this dysfunction, including forward head, rounded shoulders, and scapular dyskinesis, contribute to upper-back and shoulder pain, particularly among office workers who spend long periods of the workday on a computer. Upper crossed syndrome is a significant contributor to both neck pain and shoulder pain among computer users, which have been rated at 55–69%, and 15–52%, respectively. Despite its prevalence, knowledge about UCS and its treatment remains spotty among primary care physicians. In addition, improvements in workstation ergonomics along with hourly work breaks may be considered as primary prevention strategies for UCS. Objectives: This narrative review examines and synthesizes evidence about the epidemiology and diagnosis of UCS, along with clinical recommendations for physiotherapeutic approaches to treatment. Ergonomic measures in the workplace, including changes in the design of computer workstations so that both the keyboard and monitor are at the proper heights to minimize the risk of long-term musculoskeletal disorders, are also critical. Methods: The first author, a Doctor of Behavioral Health, performed the initial literature search, which was reviewed by the second author, a PhD in sports injury epidemiology. The third author, a chiropractor and practice owner, provided clinical recommendations for stretching and strengthening exercises, which were also described in the literature. Discussion: While easily treatable when caught early, UCS may become resistant to noninvasive approaches over time, and more severe pathologies of the neck and shoulder, including impingement, thoracic outlet syndrome, and cervicogenic headaches may result. Because there is no specific ICD code for UCS, it is important for physicians to recognize the early signs, consider them in the context of workplace-related injuries, and understand physiotherapeutic strategies for symptom resolution. Full article

Accurate trajectory tracking remains a central challenge in differential-drive mobile robots (DDMRs), particularly when operating under real-world conditions. Model Predictive Control (MPC) provides a powerful framework for this task, but its performance degrades when the robot deviates significantly from the nominal path. To address this limitation, robust recovery mechanisms are required to ensure stable and precise tracking. This work presents an experimental validation of an MPC controller applied to a four-wheel DDMR, whose odometry is corrected by a SLAM algorithm running in ROS 2. The MPC is formulated as a quadratic program with state and input constraints on linear (v) and angular ($\omega$) velocities, using a prediction horizon of $N_p=15$ future states, adjusted to the computational resources of the onboard computer. A novel dynamic look-ahead re-entry strategy is proposed, which activates when the robot exits a predefined lateral error band ($\delta =0.05$ m) and interpolates a smooth reconnection trajectory based on a forward look-ahead point, ensuring gradual convergence and avoiding abrupt re-entry actions. Accuracy was evaluated through lateral and heading errors measured via geometric projection onto the nominal path, ensuring fair comparison. From these errors, RMSE, MAE, $P_{95}$, and in-band percentage were computed as quantitative metrics. The framework was tested on real hardware at 50 Hz through 5 nominal experiments and 3 perturbed experiments. Perturbations consisted of externally imposed velocity commands at specific points along the path, while configuration parameters were systematically varied across trials, including the weight R, smoothing distance $L_{\mathrm{smooth}}$, and activation of the re-entry strategy. In nominal conditions, the best configuration (ID 2) achieved a lateral RMSE of 0.05 m, a heading RMSE of 0.06 rad, and maintained 68.8% of the trajectory within the validation band. Under perturbations, the proposed strategy substantially improved robustness. For instance, in experiment ID 6 the robot sustained a lateral RMSE of 0.12 m and preserved 51.4% in-band, outperforming MPC without re-entry, which suffered from larger deviations and slower recoveries. The results confirm that integrating MPC with the proposed re-entry strategy enhances both accuracy and robustness in DDMR trajectory tracking. By combining predictive control with a spatially grounded recovery mechanism, the approach ensures consistent performance in challenging scenarios, underscoring its relevance for reliable mobile robot navigation in uncertain environments. Full article

Navigation errors due to drifting in inertial systems using low-cost sensors are some of the main challenges for land vehicle navigation in Global Navigation Satellite System (GNSS)-denied environments. In this paper, we propose an autonomous navigation strategy with a wheel-mounted microelectromechanical system (MEMS) inertial measurement unit (IMU), referred to as the wheeled inertial navigation system (INS), to effectively suppress drifted navigation errors. The position, velocity, and attitude (PVA) of the vehicle are predicted through the inertial mechanization algorithm, while gyro outputs are utilized to derive the vehicle’s forward velocity, which is treated as an observation with non-holonomic constraints (NHCs) to estimate the inertial navigation error states. To establish a theoretical foundation for wheeled INS error characteristics, a comprehensive system observability analysis is conducted from an analytical point of view. The wheel rotation significantly improves the observability of gyro errors perpendicular to the rotation axis, which effectively suppresses azimuth errors, horizontal velocity, and position errors. This leads to the superior navigation performance of a wheeled INS over the traditional odometer (OD)/NHC/INS. Moreover, a hybrid extended particle filter (EPF), which fuses the extended Kalman filter (EKF) and PF, is proposed to update the vehicle’s navigation states. It has the advantages of (1) dealing with the system’s non-linearity and non-Gaussian noises, and (2) simultaneously achieving both a high level of accuracy in its estimation and tolerable computational complexity. Kinematic field test results indicate that the proposed wheeled INS is able to provide an accurate navigation solution in GNSS-denied environments. When a total distance of over 26 km is traveled, the maximum position drift rate is only 0.47% and the root mean square (RMS) of the heading error is 1.13°. Full article

Vehicular ad hoc networks (VANETs) must simultaneously satisfy stringent reliability, latency, and sustainability targets under highly dynamic urban and highway mobility. Existing solutions typically optimise one or two dimensions (link stability, clustering, or energy) but lack an integrated, adaptive mechanism that fuses heterogeneous metrics while remaining lightweight and deployable. This paper introduces a VANET routing protocol named SYMPHONY (Synergistic Hierarchical Metric-Fusion and Predictive Hybrid Optimization for Network Yield) that operates in three coordinated layers: (i) a compact neighbourhood filtering stage that reduces forwarding scope and eliminates transient relays, (ii) a cluster layer that elects resilient cluster heads using fuzzy energy-aware metrics and backup leadership, and (iii) a global inter-cluster optimizer that blends a GA-reseeded swarm metaheuristic with a stability-aware pheromone scheme to produce multi-objective routes. Crucially, SYMPHONY employs an ultra-lightweight online weight-adaptation module (contextual linear bandit) to tune metric fusion weights in response to observed rewards (packet delivery ratio, end-to-end delay, and Green Performance Index). We evaluated the proposed routing protocol SYMPHONY versus strong modern baselines across urban and highway scenarios with varying density and resource constraints. The results demonstrate that SYMPHONY improves packet delivery ratio by up to 12–18%, reduces latency by 20–35%, and increases the Green Performance Index by 22–45% relative to the best baseline, while keeping control overhead and per-node computation within practical bounds. Full article

Background: This study investigated how static postural parameters influence dynamic spinopelvic balance across varying walking speeds. One hundred healthy young adults (aged 18–25) underwent rasterstereographic assessment (DIERS 4Dmotion^®) to quantify static global alignment metrics including craniovertebral angle (CVA), Q-angle, sagittal and coronal imbalance, pelvic rotation, torsion, obliquity, vertebral rotation, thoracic kyphosis, lumbar lordosis, and pelvic tilt, followed by dynamic spinopelvic analysis during treadmill walking at 1, 2, 4, and 5 km/h. Methods: Multiple linear regression models were used to determine the predictive value of static postural measures for dynamic outcomes at each speed. At slower walking speeds (1–2 km/h), static alignment variables significantly predicted dynamic spinopelvic parameters (adjusted R² = 0.53–0.73; RMSE = 0.59–0.81), with CVA, sagittal imbalance, and pelvic torsion emerging as the most consistent predictors. Results: At higher speeds (4–5 km/h), predictive strength declined substantially (adjusted R² = 0.04–0.34), indicating a shift from posture-driven to neuromuscular-governed gait control. The Q-angle showed limited and inconsistent predictive value across all conditions. Conclusions: Overall, static postural alignment, particularly CVA, sagittal imbalance, and pelvic torsion, serves as a moderate predictor of spinopelvic dynamics at slow to moderate gait speeds but loses explanatory power as velocity increases, emphasizing the growing role of neuromuscular control in maintaining dynamic balance. These findings highlight the clinical relevance of integrating both static and dynamic assessments to comprehensively evaluate postural and locomotor function. Full article

Antisense genes (usually suffixed by -AS) represent a class of long non-coding RNAs (lncRNAs) transcribed from the opposite strand of annotated human genes or exon(s). A total of ~2236 human antisense genes exist in the human genome. Their genomic locations with respect to the corresponding sense genes, their dysregulated expression patterns in cancer specimens, and clinical associations with patient outcomes reveal their potential importance in clinical settings. As of today, there lacks a comprehensive review of HNC-associated antisense genes/transcripts to help move forward the antisense field for genetic biomarker development or future drug research. In total, 2.3% (52/2236 antisense genes) of all known human antisense genes have been investigated in head and neck cancer (HNC). Thus, we perform a comprehensive review of the genomic aberrations (mutations, copy number changes, RNA-expression dysregulation, and single nucleotide polymorphisms) associated with HNC patient prognosis, disease progression, cancer cell signaling, drug sensitivity, and radio-resistance. Four antisense genes, namely HOXA10-AS, LEF1-AS1, MSC-AS1, and ZEB2-AS1, have been clinically cross-validated and have consistently demonstrated to be associated with patient outcomes in multiple independent cohorts by different research teams, with clear evidence for the prioritization of clinical biomarker development in HNC. Single nucleotide polymorphisms (SNPs) of antisense genes with evidence for HNC risk or outcomes should be further validated in different ethnic groups, for potential global HNC applications. Full article

Objectives: Strong cervical musculature is recognized as a protective factor against sports-related concussions. Evidence suggests that jaw clenching may activate cervical muscles, potentially reducing head acceleration during impact. Methods: This observational cohort study examined the immediate effects of a customized interocclusal orthotic (CIO) on cervical muscle performance. Forty-two healthy adults (≥18 years) underwent strength and endurance testing with and without a CIO using a digital pressure gauge and six directional isometric contractions. Descriptive statistics and two-way repeated-measures MANOVA models were applied to evaluate condition effects. Results: CIO use produced significant improvements in cervical muscle strength and endurance across all directions compared to non-use. Forward flexion strength increased by 12.96% (p < 0.001, ηp² = 0.185), backward extension by 10.34% (p = 0.017, ηp² = 0.091), right rotation by 19.03% (p < 0.001, ηp² = 0.333) and left rotation by 19.86% (p < 0.001, ηp² = 0.353). Endurance gains demonstrated large effect sizes, with flexor endurance improving by 44.57% (p < 0.001, ηp² = 0.447). Conclusions: Optimized jaw alignment using a customized orthotic can elicit immediate, clinically meaningful enhancements in cervical strength and endurance, suggesting a promising adjunct for concussion risk mitigation in contact sports. Full article

Background/Objective: Postural control differs between individuals with lower limb amputation and the general population. Although previous studies examined the effects of unexpected surface perturbations on postural control in individuals with transtibial amputation (TTA) and individuals with transfemoral amputation (TFA), their impact on lower limb muscle activation remains unclear. This study aimed to assess postural control on a stable surface and to evaluate the effects of unexpected surface perturbations on lower limb muscle activation in unilateral TTAs, TFAs, and in a healthy control group (CG). Methods: The study included 10 TTAs, 9 TFAs, and 10 healthy controls. Postural control was assessed using a force platform, and lower limb muscle activity was recorded with surface electromyography during unexpected surface perturbations. Results: The TFAs showed the highest anteroposterior and lateral postural sway under compliant surface eyes closed and the highest lateral sway under normal surface eyes closed, whereas the CG showed the lowest values (p < 0.05). During forward perturbations, rectus femoris (RF) and tibialis anterior (TA) activations were significantly higher than biceps femoris (BF) and medial head of the gastrocnemius (GM) activations, respectively, across all groups (p < 0.05). During backward perturbations, GM activations exceeded TA activations in all groups, while BF activations were higher than RF only in TTAs (p < 0.05). Significant group effects were found for RF and BF during forward perturbations, and side effects for BF (forward) and RF (backward) activations (p < 0.05). Conclusions: Postural control responses vary with the level of lower limb amputation. TFAs relied more on visual input during quiet standing, whereas TTAs demonstrated greater reliance on thigh muscle activation during surface perturbations. These findings highlight the need to consider amputation level in balance and rehabilitation programs. Full article

Head posture assessment commonly involves measuring cranial angles, with photogrammetry favored for its simplicity over CT scans or goniometers. However, most photo-based measurements remain manual, making them time-consuming and inefficient. Existing automatic measuring approaches often requires specific markers and clean backgrounds, limiting their usability. We present AutoMCA, a robust automatic measurement system for cranial angles using accessible markers and tolerating typical indoor backgrounds. AutoMCA integrates MediaPipe Pose, a machine-learning solution, for head–neck segmentation and applies color thresholding and morphological operations for marker detection. Validation tests demonstrated Pearson correlation coefficients above 0.98 compared to manual Kinovea measurements for both the craniovertebral angle (CVA) and cranial rotation angle (CRA), confirming high accuracy. Further validation on individuals with neck disorders showed similarly strong correlations, supporting clinical applicability. Speed comparison tests revealed that AutoMCA significantly reduces measurement time compared to traditional photogrammetry. Robustness tests confirmed reliable performance across varied backgrounds and marker types. In conclusion, AutoMCA measures head posture efficiency and lowers the requirements for instruments and space, making the assessment more versatile and applicable. Full article

Accurate estimation of lithium-ion battery (LIB) state of health (SOH) is critical for prolonging battery life and ensuring safe operation. To address the limitations of existing data-driven models in robustness and feature coupling, this paper presents a new Bagging-PiFormer framework for SOH estimation. The framework integrates ensemble learning with an improved Transformer architecture to achieve accurate and stable performance across various degradation conditions. Specifically, multiple PiFormer base models are trained independently under the Bagging strategy to enhance generalization. Each PiFormer consists of a stack of PiFormer layers, which combines a cross-channel attention mechanism to model voltage–current interactions and a local convolutional feed-forward network (LocalConvFFN) to extract local degradation patterns from charging curves. Residual connections and layer normalization stabilize gradient propagation in deep layers, while a purely linear output head enables precise regression of the continuous SOH values. Experimental results on three datasets demonstrate that the proposed method achieves the lowest MAE, RMSE, and MAXE values among all compared models, reducing overall error by 10–33% relative to mainstream deep-learning methods such as Transformer, CNN-LSTM, and GCN-BiLSTM. These results confirm that the Bagging-PiFormer framework significantly improves both the accuracy and robustness of battery SOH estimation. Full article

Background/Objectives: Evaluating glenoid changes in rotator cuff tear arthropathy (RCTA) is crucial for preoperative planning. MRI with zero echo time (ZTE) sequence, which produces CT-like images, allows for the assessment of osseous morphology as well as factors contributing to pseudoparesis in RCTA patients. Methods: In this retrospective study, using 3T MRI, glenoid version, glenoid vault depth, humeral subluxation index, humeral head medialization, critical shoulder angle, glenoid best-fit circle width, glenoid best-fit circle bone loss ratio (GBLR), and anterior, central, and posterior glenoid bone loss were measured on reformatted 3D ZTE images in 43 shoulders independently by three observers. The same measurements were repeated by one observer after 10 days. Muscle cross-sectional areas were measured. Patients’ active ROMs, American Shoulder and Elbow Surgeons (ASES), and Constant–Murley scores were recorded. Patients unable to perform 90° forward elevation were classified as the pseudoparesis group. Results: Interobserver agreements were good to excellent, except for glenoid vault depth, anterior bone loss, and GBLR. Intraobserver agreements were good to excellent. The pseudoparesis group showed significantly less subscapularis muscle cross-sectional area (p = 0.006). Moderate correlations were found between subscapularis cross-sectional area and forward elevation, abduction, and internal rotation ([r = 0.471, p = 0.001]; [r = 0.447, p = 0.003]; [r = 0.464, p = 0.002], respectively). Moderate negative correlations were found between anterior glenoid loss and forward elevation (r = −0.411, p = 0.006) and abduction (r = −0.475, p = 0.001). Conclusions: MRI with ZTE sequence demonstrated good reliability for assessing osseous morphology in shoulders with RCTA. Glenoid anterior bone loss and loss of subscapularis muscle are both associated with pseudoparesis. Full article

Forward-looking sonar (FLS) image segmentation is essential for underwater exploration with remaining challenges including low contrast, ambient noise, and complex backgrounds, which both existing traditional and deep learning-based methods fail to address effectively. This paper presents TriEncoderNet, a novel model that simultaneously extracts local, global, and edge-related features through three parallel encoders. Specifically, the model integrates a convolutional neural network (CNN) for local feature extraction, a transformer for global context modeling, and a histogram of oriented gradients (HOG) encoder for edge and shape detection. The key innovations of TriEncoderNet include the CrossFusionTransformer (CFT) module, which effectively integrates local and global features to capture both fine details and comprehensive context, and the HOG attention gate (HAG) module, which enhances edge detection and preserves semantic consistency across diverse feature types. Additionally, TriEncoderNet introduces the hierarchical efficient transformer (HETransformer) with a lightweight multi-head self-attention mechanism to reduce computational overhead while maintaining global context modeling capability. Experimental results on the marine debris dataset and UATD dataset demonstrate the superior performance of TriEncoderNet. Specifically, it achieves an mIoU of 0.793 and mAP of 0.916 on the marine debris dataset, and an mIoU of 0.582 and mAP of 0.687 on the UATD Dataset, outperforming state-of-the-art methods in both segmentation accuracy and robustness in challenging underwater environments. Full article

Traditional calibration methods rely on precise targets and frequent manual intervention, making them time-consuming and unsuitable for large-scale deployment. Existing learning-based approaches, while automating the process, are typically limited to single LiDAR–camera pairs, resulting in poor scalability and high computational overhead. To address these limitations, we propose a lightweight calibration network with flexibility in the number of sensor pairs, making it capable of jointly calibrating multiple cameras and LiDARs in a single forward pass. Our method employs a frozen pre-trained Swin Transformer as a shared backbone to extract unified features from both RGB images and corresponding depth maps. Additionally, we introduce a cross-modal channel-wise attention module to enhance key feature alignment and suppress irrelevant noise. Moreover, to handle variations in viewpoint, we design a modular calibration head that independently estimates the extrinsics for each LiDAR–camera pair. Through large-scale experiments on the nuScenes dataset, we show that our model, requiring merely 78.79 M parameters, attains a mean translation error of 2.651 cm and a rotation error of $0.{246}^{\circ }$, achieving comparable performance to existing methods while significantly reducing the computational cost. Full article

Energy efficiency and prolonged network lifetime remain central challenges in wireless sensor networks (WSNs). Clustering, cluster-head (CH) selection, and routing are key to addressing these issues because they directly affect energy consumption, data delivery, and overall network stability. This paper introduces a novel hybrid protocol, PUMA-GRID, which integrates the recently proposed Puma Optimization Algorithm with a grid-based multi-hop routing framework. Unlike traditional schemes, PUMA-GRID adaptively balances exploration and exploitation during CH selection while learning energy-efficient data-forwarding paths through grid-based routing. This combination improves adaptability, scalability, and load balancing, which distinguishes PUMA-GRID from the primary metaheuristic competitor AEO-GRID, as well as earlier AEO, LEACH, and static PUMA variants. The fitness function for CH election incorporates intra-cluster distance, distance to the base station (BS), and residual energy, with adjustable weights that enable flexible adaptation to different deployment scenarios. Simulation experiments were performed under various BS placements and weight configurations to assess the influence of each factor. The results show that the impact of the weights depends strongly on BS location and that careful tuning is required to balance efficiency and fairness. Across all scenarios, PUMA-GRID demonstrates superior performance compared with LEACH, AEO-based schemes, and other PUMA variants. Overall, PUMA-GRID provides an effective and scalable solution for sustainable, energy-aware operation in WSNs. Full article