Search Results (66,359)

Hyperspectral image (HSI) reconstruction is a critical and indispensable step in spectral compressive imaging (CASSI) systems and directly affects our ability to capture high-quality images in dynamic environments. Recent research has increasingly focused on deep unfolding frameworks for HSI reconstruction, showing notable progress. However, these approaches have to break the optimization task into two sub-problems, solving them iteratively over multiple stages, which leads to large models and high computational overheads. This study presents a simple yet effective method that passes the degradation information (sensing mask) through a deep learning network to disentangle the degradation and the latent target’s representations. Specifically, we design a lightweight MLP block to capture non-local similarities and long-range dependencies across both spatial and spectral domains, and investigate an attention-based mask modelling module to achieve the spatial–spectral-adaptive degradation representationthat is fed to the MLP-based network. To enhance the information flow between MLP blocks, we introduce a multi-level fusion module and apply reconstruction heads to different MLP features for deeper supervision. Additionally, we combine the projection loss from compressive measurements with reconstruction loss to create a dual-domain loss, ensuring consistent optical detection during HS reconstruction. Experiments on benchmark HS datasets show that our method outperforms state-of-the-art approaches in terms of both reconstruction accuracy and efficiency, reducing computational and memory costs. Full article

Place recognition plays a crucial role in tasks such as loop closure detection and re-localization in robotic navigation. As a high-level representation within scenes, semantics enables models to effectively distinguish geometrically similar places, therefore enhancing their robustness to environmental changes. Unlike most existing semantic-based LiDAR place recognition (LPR) methods that adopt a multi-stage and relatively segregated data-processing and storage pipeline, we propose a novel end-to-end LPR model guided by semantic information—SG-LPR. This model introduces a semantic segmentation auxiliary task to guide the model in autonomously capturing high-level semantic information from the scene, implicitly integrating these features into the main LPR task, thus providing a unified framework of “segmentation-while-describing” and avoiding additional intermediate data-processing and storage steps. Moreover, the semantic segmentation auxiliary task operates only during model training, therefore not adding any time overhead during the testing phase. The model also combines the advantages of Swin Transformer and U-Net to address the shortcomings of current semantic-based LPR methods in capturing global contextual information and extracting fine-grained features. Extensive experiments conducted on multiple sequences from the KITTI and NCLT datasets validate the effectiveness, robustness, and generalization ability of our proposed method. Our approach achieves notable performance improvements over state-of-the-art methods. Full article

As mining technology advances, intelligent robots in open-pit mining require precise localization and digital maps. Nonetheless, significant pitch variations, uneven highways, and rocky surfaces with minimal texture present substantial challenges to the precision of feature extraction and positioning in traditional visual SLAM systems, owing to the intricate terrain features of open-pit mines. This study proposes an improved SLAM technique that integrates visual and Inertial Measurement Unit (IMU) data to address these challenges. The method incorporates a point–line feature fusion matching strategy to enhance the quality and stability of line feature extraction. It integrates an enhanced Line Segment Detection (LSD) algorithm with short segment culling and approximate line merging techniques. The combination of IMU pre-integration and visual feature restrictions is executed inside a tightly coupled visual–inertial framework utilizing a sliding window approach for back-end optimization, enhancing system robustness and precision. Experimental results demonstrate that the suggested method improves RMSE accuracy by 36.62% and 26.88% on the MH and VR sequences of the EuRoC dataset, respectively, compared to ORB-SLAM3. The improved SLAM system significantly reduces trajectory drift in the simulated open-pit mining tests, improving localization accuracy by 40.62% and 61.32%. The results indicate that the proposed method demonstrates significance. Full article

With advancements in autonomous driving, LiDAR has become central to 3D object detection due to its precision and interference resistance. However, challenges such as point cloud sparsity and unstructured data persist. This study introduces MS3D (Multi-Scale Feature Fusion 3D Object Detection Method), a novel approach to 3D object detection that leverages the architecture of a 2D Convolutional Neural Network (CNN) as its core framework. It integrates a Second Feature Pyramid Network to enhance multi-scale feature representation and contextual integration. The Adam optimizer is employed for efficient adaptive parameter tuning, significantly improving detection performance. On the KITTI dataset, MS3D achieves average precisions of 93.58%, 90.91%, and 88.46% in easy, moderate, and hard scenarios, respectively, surpassing state-of-the-art models like VoxelNet, SECOND, and PointPillars. Full article

Background: In this study, we aimed to investigate the relationship between diabetic nephropathy (DN) and triglyceride/glucose (TyG) index and triglyceride/high-density lipoprotein cholesterol ratio (Tg/HDL-C) as surrogate markers of insulin resistance. Method: Medical records of 15,378 individuals between February 2019 and May 2024 were examined. Serum glucose, Tg, HDL-C, HbA1c, estimated glomerular filtration rate (eGFR), and urine albumin/creatinine ratio (UACR) were evaluated and the TyG index and TG/HDL-C ratios were calculated for each individual. DN was defined as a UACR ≥ 30 mg/g and/or eGFR <60 mL/min/1.73 m². Results: Of 10,714 patients, DN was detected in 3763 (35.1%). Females had 10% higher odds of developing DN compared to males. A TyG index at or above the determined cutoff point (9.58) indicated a risk of DN and the sensitivity and specificity values were 44.01% and 71.28%, respectively. The risk of DN was 1.95 times higher in individuals with a TyG index value of ≥9.58 compared to those with a TyG index <9.58. While the Tg/HDL ratio was significant in detecting DN in the univariate analysis (odds ratio (OR) 1.59; 95% confidence interval 1.46–1.73), this significance was not found in the multivariate analysis (OR 1.15; 95% confidence interval 0.94–1.40). Conclusion: A high TyG index is associated with DN in patients with type 2 diabetes and it might be a potential marker in predicting DN. Full article

With a view to improve measurements, this paper presents a statistical approach for characterizing the behaviour of roughness parameters based on measurements performed on ground surface topographies (grit #080/#120). A S neox^TM (Sensofar^®, Terrassa, Spain), equipped with three optical instrument modes (Focus Variation (FV), Coherence Scanning Interferometry (CSI), and Confocal Microscopy (CM)), is used according to a specific measurement plan, called Morphomeca Monitoring, including topography representativeness and several time-based measurements. Previously applied to the Sa parameter, the statistical approach based here solely on the Quality Index (QI) has now been extended to a multi-parameter approach. Firstly, the study focuses on detecting and explaining parameter disturbances in raw data by identifying and quantifying outliers of the parameter’s values, as a new first indicator. This allows us to draw parallels between these outliers and the surface topography, providing reflection tracks. Secondly, the statistical approach is applied to highlight disturbed parameters concerning the instrument mode used and the concerned grit level with two other indicators computed from QI, named homogeneity and number of modes. The applied method shows that a cleaning of the data containing the parameters values is necessary to remove outlier values, and a set of roughness parameters could be determined according to the assessment of the indicators. The final aim is to provide a set of parameters which best describe the measurement conditions based on monitoring data, statistical indexes, and surface topographies. It is shown that the parameters Sal, Sz and Sci are the most reliable roughness parameters, unlike Sdq and S5p, which appear as the most unstable parameters. More globally, the volume roughness parameters appear as the most stable, differing from the form parameters. This investigated point of view offers thus a complementary framework for improving measurement processes. In addition, this method aims to provide a global and more generalizable alternative than traditional methods of uncertainty calculation, based on a thorough analysis of multi-parameter and statistical indexes. Full article

Background/Objectives: Cutaneous infections of fungal and bacterial origins are common. An accurate diagnosis—especially concerning pathogens that are difficult to isolate on culture—can be achieved using molecular methods (PCR) with a short turnaround time. Methods: We reviewed records of skin specimens (superficial scrapings) submitted by dermatologists across the United States with a clinically suspected dermatitis. As per physician’s order, specimens were tested for infections either fungal (N = 4262) or bacterial (N = 1707) in origin. All unique specimens (one per patient) were subjected to real-time PCR assays where cases suspected of a fungal etiology were tested for dermatophytes, Malassezia and Candida, and cases suspected of a bacterial etiology were tested for Streptococcus pyogenes, Staphylococcus aureus, and the mecA gene potentially conferring β-lactam resistance. Results: Fungal agents were detected in 32.8% (SD: 4.5) of the submitted specimens, with most attributed to dermatophytes (19.3% (SD: 4.9)), followed by Malassezia (8.7% (SD: 2.8)) and Candida (2.9% (SD: 1.0)). Dermatophyte detection was more common in the elderly (≥65 years) compared to young adults (18–44 years) (OR: 1.8 (95% CI: 1.5, 2.2)), whereas Malassezia was more commonly detected in younger age groups (12.1–13.6%) than the elderly (5.6%). Candida was more frequently observed in females while dermatophytes and Malassezia were more frequently observed in males. Approximately one quarter of the submitted skin specimens tested positive for S. aureus (23.6% (SD: 3.4)), of which 34.4% (SD: 9.8) exhibited concurrent detection of the mecA gene. An S. aureus detection was more frequently observed in males (OR: 1.5 (95% CI: 1.2, 1.9)) and in children (OR: 1.7 (95% CI: 1.2, 2.5)). Streptococcus pyogenes was rarely detected. Among specimens positive for dermatophytes, 12.0% (20/166) showed co-detection of S. aureus and mecA, which is in contrast to 6.8% (70/1023) detected in samples without a fungal co-detection and 6.2% (8/130) in samples positive for Malassezia. Conclusions: PCR testing, when available, can be valuable as a part of routine care for diagnosing patients with clinically suspected skin infections. Further studies are warranted to survey the prevalence of resistant S. aureus isolates in dermatology outpatients, in particular with regard to the association with dermatophyte infections. Full article

In applications such as aerial object interception and ballistic estimation, it is crucial to precisely detect the centroid position of the target rather than to merely identify the position of the target bounding box or segment all pixels belonging to the target. Due to the typically long distances between targets and imaging devices in such scenarios, targets often exhibit a low contrast and appear as dim, obscure shapes in infrared images, which represents a challenge for human observation. To rapidly and accurately detect small targets, this paper proposes a lightweight, end-to-end detection network for small infrared targets. Unlike existing methods, the input of this network is five consecutive images after background matching. This design significantly improves the network’s ability to extract target motion features and effectively reduces the interference of static backgrounds. The network mainly consists of a local feature aggregation module (LFAM), which uses multiple-sized convolution kernels to capture multi-scale features in parallel and integrates multiple spatial attention mechanisms to achieve accurate feature fusion and effective background suppression, thereby enhancing the ability to detect small targets. To improve the accuracy of predicted target centroids, a centroid correction algorithm is designed. In summary, this paper presents a lightweight centroid detection network based on background matching for weak, small infrared targets. The experimental results show that, compared to directly inputting a sequence of images into the neural network, inputting a sequence of images processed by background matching can increase the detection rate by 9.88%. Using the centroid correction algorithm proposed in this paper can therefore improve the centroid localization accuracy by 0.0134. Full article

Aim: The main purpose of the present study was to investigate the carrier rate of Aggregatibacter actinomycetemcomitans and its JP2 genotype in a cohort of 200 Western Australians diagnosed with periodontitis. Materials and Methods: In this descriptive cross-sectional study, 220 consecutive patients with periodontitis, aged 18 years and older, were recruited to a specialist periodontal practice in Perth City. Every patient included in this study contributed three different intra-oral samples. Periodontal, radiographical, and microbiological assessments were performed. The samples were analysed using a polymerase chain reaction for the detection of Aggregatibacter actinomycetemcomitans and its JP2 genotype using the primers and conditions described previously. A Chi-square test and logistic regression analysis were performed to evaluate the results. Results: The prevalence of Aggregatibacter actinomycetemcomitans was 28.18%. The carrier rates of A. actinomycetemcomitans in the unstimulated saliva, cheek swabs, and pooled subgingival plaque samples were 21.80%, 19.50%, and 17.70%, respectively. There was a significant correlation between the severe form of periodontitis (stage IV, grade C) and younger age (p = 0.004), positive family history of periodontitis (p < 0.001), oral hygiene method (p < 0.001), and irregular dental visit attendance (p < 0.001). The binary logistic regression analysis revealed that having severe periodontitis risk increased almost three times in those who were young (OR: 2.812) and came from a family with a history of periodontal disease (OR: 3.194). However, the risk of severe periodontitis was five times higher in those patients with tooth loss due to periodontal disease (OR: 5.071). The highly leukotoxic JP2 genotype of Aggregatibacter actinomycetemcomitans was not detected. Conclusions: This study of a Western Australian cohort confirmed the low presence of Aggregatibacter actinomycetemcomitans and the complete absence of its JP2 genotype. Young age, family history of periodontal disease, lack of flossing, irregular dental visits, and tooth loss due to periodontitis were identified as potential risk factors for periodontitis stage IV, grade C in this cohort. Full article

Infrared image segmentation in marine environments is crucial for enhancing nighttime observations and ensuring maritime safety. While recent advancements in deep learning have significantly improved segmentation accuracy, challenges remain due to nighttime marine scenes including low contrast and noise backgrounds. This paper introduces a cross-granularity infrared image segmentation network CGSegNet designed to address these challenges specifically for infrared images. The proposed method designs a hybrid feature framework with cross-granularity to enhance segmentation performance in complex water surface scenarios. To suppress feature semantic disparity against different feature granularity, we propose an adaptive multi-scale fusion module (AMF) that combines local granularity extraction with global context granularity. Additionally, incorporating a handcrafted histogram of oriented gradients (HOG) features, we designed a novel HOG feature fusion module to improve edge detection accuracy under low-contrast conditions. Comprehensive experiments conducted on the public infrared segmentation dataset demonstrate that our method outperforms state-of-the-art techniques, achieving superior segmentation results compared to professional infrared image segmentation methods. The results highlight the potential of our approach in facilitating accurate infrared image segmentation for nighttime marine observation, with implications for maritime safety and environmental monitoring. Full article

This work describes the development of an electrochemical biosensor method based on bacterial consortia to determine antioxidant capacity. The bacterial consortium used is a combination of bacteria from the genera Bacillus and Pseudomonas which can produce the enzymes tyrosinase and laccase. The consortium bacteria were immobilized on the surface of the screen-printed carbon electrode (SPCE) to form a biofilm. Biofilms were selected based on the highest current response evaluated electrochemically using cyclic voltammetry analysis techniques. Optimum consortium biofilm conditions were obtained in a phosphate buffer solution of pH 7, and biofilm formation occurred on day 7. This work produces analytical performance with a coefficient of determination (R²) of 0.9924. The limit of detection (LOD) and limit of quantification (LOQ) values are 0.5 µM and 10 µM, respectively. The biosensor showed a stable response until the 10th week. This biosensor was used to measure the antioxidant capacity of five extracts, and the results were confirmed using a standard method, the 2,2-diphenyl-1-picrylhydrazyl (DPPH) method. The highest antioxidant capacity is guava extract and the lowest is tempuyung extract. Thus, the development of this biosensor method can be used as an alternative for measuring antioxidant capacity. Full article

Background: The increasing integration of AI in chest X-ray evaluation holds promise for enhancing diagnostic accuracy and optimizing clinical workflows. However, understanding its performance in real-world clinical settings is essential. Objectives: In this study, we evaluated the sensitivity (Se) and specificity (Sp) of an AI-based software (Arterys MICA v29.4.0) alongside a radiology resident in interpreting chest X-rays referred from the emergency department (ED), using a senior radiologist’s assessment as the gold standard (GS). We assessed the concordance between the AI system and the resident, noted the frequency of doubtful cases for each category, identified how many were considered positive by the GS, and assessed variables that AI was not trained to detect. Methods: We conducted a retrospective observational study analyzing chest X-rays from a sample of 784 patients referred from the ED at our hospital. The AI system was trained to detect five categorical variables—pulmonary nodule, pulmonary opacity, pleural effusion, pneumothorax, and fracture—and assign each a confidence label (“positive”, “doubtful”, or “negative”). Results: Sensitivity in detecting fractures and pneumothorax was high (100%) for both AI and the resident, moderate for pulmonary opacity (AI = 76%, resident = 71%), and acceptable for pleural effusion (AI = 60%, resident = 67%), with negative predictive values (NPV) above 95% and areas under the curve (AUC) exceeding 0.8. The resident showed moderate sensitivity (75%) for pulmonary nodules, while AI’s sensitivity was low (33%). AI assigned a “doubtful” label to some diagnoses, most of which were deemed negative by the GS; the resident expressed doubt less frequently. The Kappa coefficient between the resident and AI was fair (0.3) across most categories, except for pleural effusion, where concordance was moderate (0.5). Our study highlighted additional findings not detected by AI, including 16% prevalence of mediastinal abnormalities, 20% surgical materials, and 20% other pulmonary findings. Conclusions: Although AI demonstrated utility in identifying most primary findings—except for pulmonary nodules—its high NPV suggests it may be valuable for screening. Further training of the AI software and broadening its scope to identify additional findings could enhance its detection capabilities and increase its applicability in clinical practice. Full article

The traditional scoliosis examination based on X-ray film is not suitable for large-scale screening, and it is also not suitable for dynamic evaluation during rehabilitation. Therefore, based on computer vision technology, this paper puts forward an evaluation method of scoliosis with different photos of the back taken by mobile phones, which involves three aspects: first, based on the key point detection model of YOLOv8, an algorithm for judging the type of spinal coronal curvature is proposed; second, an algorithm for evaluating the coronal plane of the spine based on the key points of the human back is proposed, aiming at quantifying the deviation degree of the spine in the coronal plane; third, the measurement algorithm of trunk rotation (ATR angle) based on multi-scale automatic peak detection (AMPD) is proposed, aiming at quantifying the deviation degree of the spine in sagittal plane. The public dataset and clinical paired data (mobile phone photo and X-ray) are used to test. The results show that this method has high accuracy and effectiveness in distinguishing the type of spinal curvature and evaluating the degree of deviation, which is higher than other deep learning algorithms. Full article

The development of energy-efficient, sensitive, and reliable gas sensors for monitoring NO₂ concentrations has garnered considerable attention in recent years. In this manuscript, TiO₂ nanotube arrays/reduced graphene oxide nanocomposites with varying rGO contents (TiO₂ NTs/rGO) were synthesized via a two-step method for room temperature NO₂ gas detection. From SEM and TEM images, it is evident that the rGO sheets not only partially surround the TiO₂ nanotubes but also establish interconnection bridges between adjacent nanotubes, which is anticipated to enhance electron–hole separation by facilitating electron transfer. The optimized TiO₂ NTs/rGO sensor demonstrated a sensitive response of 19.1 to 1 ppm of NO₂, a 5.26-fold improvement over the undoped TiO₂ sensor. Additionally, rGO doping significantly enhanced the sensor’s response/recovery times, reducing them from 24 s/42 s to 18 s/33 s with just 1 wt.% rGO. These enhancements are attributed to the increased specific surface area, higher concentration of chemisorbed oxygen species, and the formation of p-n heterojunctions between TiO₂ and rGO within the nanocomposites. This study provides valuable insights for the development of TiO₂/graphene-based gas sensors for detecting oxidizing gases at room temperature. Full article

Electromagnetic technology is widely applied in numerous fields, and precise electromagnetic characteristic fitting technology has become a crucial part for enhancing system performance and optimizing design. However, it faces challenges such as high computational complexity and the difficulty in balancing the accuracy and generalization ability of the model. For example, the Radar Cross Section (RCS) distribution characteristics of a single corner reflector model or Luneberg lens provide a relatively stable RCS value within a certain airspace range, which to some extent reduces the difficulty of radar target detection and fails to truly evaluate the radar performance. This paper aims to propose an innovative multi-parameter optimization method for electromagnetic characteristic fitting based on deep learning. By selecting common targets such as reflectors and Luneberg lens reflectors as optimization variables, a deep neural network model is constructed and trained with a large amount of electromagnetic data to achieve high-precision fitting of the target electromagnetic characteristics. Meanwhile, an advanced genetic optimization algorithm is introduced to optimize the multiple parameters of the model to meet the error index requirements of radar target detection. In this paper, by combining specific optimization variables such as corner reflectors and Luneberg lenses with the deep learning model and genetic algorithm, the deficiencies of traditional methods in handling electromagnetic characteristic fitting are effectively addressed. The experimental results show that the 60° corner reflector successfully realizes the simulation of multiple peak characteristics of the target, and the Luneberg lens reflector achieves the simulation of a relatively small RCS average value with certain fluctuations in a large space range, which strongly proves that this method has significant advantages in improving the fitting accuracy and optimization efficiency, opening up new avenues for research and application in the electromagnetic field. Full article