Search Results (38)

With the rapid development of wireless communication technologies and the increasing demand for efficient spectrum utilization, broadband spectrum sensing has become critical in both civilian and military fields. Detecting narrowband signals under broadband environments, especially under low-signal-to-noise-ratio (SNR) conditions, poses significant challenges due to the complexity of time–frequency features and noise interference. To this end, this study presents a signal detection model named deformable feature-enhanced network–You Only Look Once (DFN-YOLO), specifically designed for blind signal detection in broadband scenarios. The DFN-YOLO model incorporates a deformable channel feature fusion network (DCFFN), replacing the concatenate-to-fusion (C2f) module to enhance the extraction and integration of channel features. The deformable attention mechanism embedded in DCFFN adaptively focuses on critical signal regions, while the loss function is optimized to the focal scaled intersection over union (Focal_SIoU), improving detection accuracy under low-SNR conditions. To support this task, a signal detection dataset is constructed and utilized to evaluate the performance of DFN-YOLO. The experimental results for broadband time–frequency spectrograms demonstrate that DFN-YOLO achieves a mean average precision (mAP50–95) of 0.850, averaged over IoU thresholds ranging from 0.50 to 0.95 with a step of 0.05, significantly outperforming mainstream object detection models such as YOLOv8, which serves as the benchmark baseline in this study. Additionally, the model maintains an average time estimation error within $5.55\times {10}^{-5}$ s and provides preliminary center frequency estimation in the broadband spectrum. These findings underscore the strong potential of DFN-YOLO for blind signal detection in broadband environments, with significant implications for both civilian and military applications. Full article

This paper examines the impacts of AI-powered assistive technologies (AIATs) on the academic success of higher education university students with visual impairments. As digital learning contexts become progressively more prevalent in higher education institutions, it is critical to understand how these technologies foster the academic success of university students with blindness or low vision. Based on the Unified Theory of Acceptance and Use of Technology (UTAUT) model, the study conducted a quantitative research approach and collected data from 390 visually impaired students who were enrolled in different universities across Saudi Arabia (SA). Employing Partial Least Squares Structural Equation Modeling (PLS-SEM), the paper tested the influences of four UTAUT dimensions—Performance Expectancy (PE), Effort Expectancy (EE), Social Influence (SI), and Facilitating Conditions (FC)—on Academic Performance (AP), while also evaluating the mediating role of Behavioral Intention (BI). The results revealed a significant positive relationship between the implementation of AI-based assistive tools and students’ academic success. Particularly, BI emerged as a key mediator in these intersections. The results indicated that PE (β = 0.137, R² = 0.745), SI (β = 0.070, R² = 0.745), and BI (β = 0.792, R² = 0.745) significantly affected AP. In contrast, EE (β = −0.041, R² = 0.745) and FC (β = −0.004, R² = 0.745) did not have a significant effect on AP. Concerning predictors of BI, PE (β = 0.412, R² = 0.317), SI (β = 0.462, R² = 0.317), and EE (β = 0.139, R² = 0.317) were all positively associated with BI. However, FC had a significant negative association with BI (β = −0.194, R² = 0.317). Additionally, the analysis revealed that EE, SI, and PE can all indirectly enhance Academic Performance by influencing BI. The findings provide practical insights for higher education policymakers, higher education administrators, and AI designers, emphasizing the need to improve the accessibility and usability of sustainable and long-term assistive technologies to better accommodate learners with visual impairments in higher education contexts. Full article

Managing traffic flow through urban intersections is challenging. Conflicts involving a mix of different vehicles with blind spots makes it relatively vulnerable for crashes to happen. This paper presents a new framework based on a fine-tuned Multimodal Large Language Model (MLLM), GPT-4o, that can control intersections using bird eye view videos taken by drones in real-time. This fine-tuned GPT-4o model is used to logically and visually reason traffic conflicts and provide instructions to the drivers, which aids in creating a safer and more efficient traffic flow. To fine-tune and evaluate the model, we labeled a dataset that includes three-month drone videos, and their corresponding trajectories recorded in Dresden, Germany, at a 4-way intersection. Preliminary results showed that the fine-tuned GPT-4o achieved an accuracy of about 77%, outperforming zero-shot baselines. However, using continuous video-frame sequences, the model performance increased to about 89% on a time serialized dataset and about 90% on an unbalanced real-world dataset, respectively. This proves the model’s robustness in different conditions. Furthermore, manual evaluation by experts includes scoring the usefulness of the predicted explanations and recommendations by the model. The model surpassed on average rating of 8.99 out of 10 for explanations, and 9.23 out of 10 for recommendations. The results demonstrate the advantages of combining MLLMs with structured prompts and temporal information for conflict detection. These results offer a flexible and robust prototype framework to improve the safety and effectiveness of uncontrolled intersections. The code and labeled dataset used in this study are publicly available (see Data Availability Statement). Full article

For the full cross-sectional profile measurement of high-aspect-ratio micro-grooves, traditional measurement methods have blind measurement areas in the vertical sidewall and its intersection area with the bottom. This paper proposes a deflection-based scanning method that utilizes a large length-to-diameter ratio probe to achieve a full cross-sectional profile measurement of micro-grooves. Blind measurement areas were eliminated by a deflection-based scanning method. The complete groove profile was obtained by stitching the positive and reversal deflection-based measurement results. The optimal deflection angle of the probe was calculated by considering the profile-stitching setting and the principle of minimizing the probe deformation during the measurement process. A four-axis measurement system was established to measure high-aspect-ratio micro-grooves, which incorporated a force feedback mechanism to maintain a constant contact force during the measurement and an integrated error separation module to modify the measurement results. The measurement method and system were experimentally validated to achieve a full cross-sectional profile measurement of micro-grooves with a width of 50 μm and an aspect ratio of no less than 3. The standard deviation of the measurement results was 82 nm, and the expanded uncertainty was 108 nm. Full article

The lithofacies of a reservoir contain key information such as rock lithology, sedimentary structures, and mineral composition. Accurate prediction of shale reservoir lithofacies is crucial for identifying sweet spots for oil and gas development. However, obtaining shale lithofacies through core sampling during drilling is challenging, and the accuracy of traditional logging curve intersection methods is insufficient. To efficiently and accurately predict shale lithofacies, this study proposes a hybrid model called Stacking, which combines four classifiers: Random Forest, HistGradient Boosting, Extreme Gradient Boosting, and Categorical Boosting. The model employs the Grid Search Method to automatically search for optimal hyperparameters, using the four classifiers as base learners. The predictions from these base learners are then used as new features, and a Logistic Regression model serves as the final meta-classifier for prediction. A total of 3323 data points were collected from six wells to train and test the model, with the final performance evaluated on two blind wells that were not involved in the training process. The results indicate that the stacking model accurately predicts shale lithofacies, achieving an Accuracy, Recall, Precision, and F1 Score of 0.9587, 0.959, 0.9587, and 0.9587, respectively, on the training set. This achievement provides technical support for reservoir evaluation and sweet spot prediction in oil and gas exploration. Full article

Data fusion is an important issue in bridge health monitoring. Through data fusion, specific unknown bridge responses can be estimated with measured responses. However, existing data fusion methods always require a precise finite element model of the bridge or partially measured target responses, which are hard to realize in actual engineering. In this study, we propose a novel data fusion method. Measured inclinations across multiple cross-sections of the target bridge and accelerations at a subset of these sections were used to estimate accelerations at the remaining sections. Theoretical analysis of a typical vehicle-bridge interaction (VBI) system has shown parallels with the blind source separation (BSS) problem. Based on this, Independent Component Analysis (ICA) was applied to derive surrogate inclination mode shapes. This was followed by calculating surrogate displacement mode shapes through numerical integration. Finally, a surrogate inter-section transfer matrix for both measured and unmeasured accelerations was constructed, enabling the estimation of the target accelerations. This paper presents three key principles involving the relationship between the surrogate and actual inter-section transfer matrices, the integration of mode shape functions, and the consistency of transfer matrices for low- and high-frequency responses, which form the basis of the proposed method. A series of numerical simulations and a large-scale laboratory experiment were proposed to validate the proposed method. Compared to existing approaches, our proposed method stands out as a purely data-driven technique, eliminating the need for finite element analysis assessment. By incorporating the ICA algorithm and surrogate mode shapes, this study addresses the challenges associated with obtaining accurate mode shape functions from low-frequency responses. Moreover, our method does not require partial measurements of the target responses, simplifying the data collection process. The validation results demonstrate the method’s practicality and convenience for real-world engineering applications, showcasing its potential for broad adoption in the field. Full article

Glaucoma, a leading cause of permanent blindness worldwide, necessitates early detection to prevent vision loss, a task that is challenging and time-consuming when performed manually. This study proposes an automatic glaucoma detection method on enhanced retinal images using deep learning. The system analyzes retinal images, generating masks for the optic disc and optic cup, and providing a classification for glaucoma diagnosis. We employ a U-Net architecture with a pretrained residual neural network (ResNet34) for segmentation and an EfficientNetB0 for classification. The proposed framework is tested on publicly available datasets, including ORIGA, REFUGE, RIM-ONE DL, and HRF. Our work evaluated the U-Net model with five pretrained backbones (ResNet34, ResNet50, VGG19, DenseNet121, and EfficientNetB0) and examined preprocessing effects. We optimized model training with limited data using transfer learning and data augmentation techniques. The segmentation model achieves a mean intersection over union (mIoU) value of 0.98. The classification model shows remarkable performance with 99.9% training and 100% testing accuracy on ORIGA, 99.9% training and 99% testing accuracy on RIM-ONE DL, and 98% training and 100% testing accuracy on HRF. The proposed model outperforms related works and demonstrates potential for accurate glaucoma classification and detection tasks. Full article

Background: The aim of this study was to compare en-face optical coherence tomography (OCT) imaging and confocal scanning laser ophthalmoscopy (cSLO) imaging at different wavelengths to identify the internal limiting membrane (ILM) peeling area after primary surgery with vitrectomy and ILM peeling for macular hole (MH). Methods: In total, 50 eyes of 50 consecutive patients who underwent primary surgery with vitrectomy and ILM peeling for MH were studied. The true ILM rhexis based on intraoperative color fundus photography was compared to the presumed ILM rhexis identified by a blinded examiner using en-face OCT imaging and cSLO images at various wavelengths. To calculate the fraction of overlap (FoO), the common intersecting area and the total of both areas were measured. Results: The FoO for the measured areas was 0.93 ± 0.03 for en-face OCT, 0.76 ± 0.06 for blue reflectance (BR; 488 nm), 0.71 ± 0.09 for green reflectance (GR; 514 nm), 0.56 ± 0.07 for infrared reflectance (IR; 815 nm) and 0.73 ± 0.06 for multispectral (MS). The FoO in the en-face OCT group was significantly higher than in all other groups, whereas the FoO in the IR group was significantly lower compared to all other groups. No significant differences were observed in FoO among the MS, BR, and GR groups. In en-face OCT, there was no significant change in the ILM peeled area measured intraoperatively and postoperatively (8.37 ± 3.01 vs. 8.24 ± 2.81 mm²; p = 0.8145). Nasal-inferior foveal displacement was observed in 38 eyes (76%). Conclusions: En-face OCT imaging demonstrates reliable postoperative visualization of the ILM peeled area. Although the size of the ILM peeling remains stable after one month, our findings indicate a notable inferior-nasal shift of the overall ILM peeling area towards the optic disc. Full article

Background/Objectives: The intermediate femoral cutaneous nerve (IFCN), the saphenous nerve, and the medial femoral cutaneous nerve (MFCN) innervate the skin of the anteromedial knee region. However, it is unknown whether the MFCN has a deeper innervation. This would be relevant for total knee arthroplasty (TKA) that intersects deeper anteromedial genicular tissue layers. Primary aim: to investigate deeper innervation of the anterior and posterior MFCN branches (MFCN-A and MFCN-P). Secondary aim: to investigate MFCN innervation of the skin covering the anteromedial knee area and medial parapatellar arthrotomy used for TKA. Methods: This study consists of (1) a dissection study and (2) unpublished data and post hoc analysis from a randomized controlled double-blinded volunteer trial (EudraCT number: 2020-004942-12). All volunteers received bilateral active IFCN blocks (nerve block round 1) and saphenous nerve blocks (nerve block round 2). In nerve block round 3, all volunteers were allocated to a selective MFCN-A block. Results: (1) The MFCN-A consistently innervated deeper structures in the anteromedial knee region in all dissected specimens. No deep innervation from the MFCN-P was observed. (2) Sixteen out of nineteen volunteers had an unanesthetized skin gap in the anteromedial knee area and eleven out of the nineteen volunteers had an unanesthetized gap on the skin covering the medial parapatellar arthrotomy before the active MFCN-A block. The anteromedial knee area and medial parapatellar arthrotomy was completely anesthetized after the MFCN-A block in 75% and 82% of cases, respectively. Conclusions: The MFCN-A shows consistent deep innervation in the anteromedial knee region and the area of MFCN-A innervation overlaps the skin area covering the medial parapatellar arthrotomy. Further trials are mandated to investigate whether an MFCN-A block translates into a clinical effect on postoperative pain after total knee arthroplasty or can be used for diagnosis and interventional pain management for chronic neuropathic pain due to damage to the MFCN-A during surgery. Full article

Private set intersection (PSI) enables two parties to determine the intersection of their respective datasets without revealing any information beyond the intersection itself. This paper particularly focuses on the scenario of unbalanced PSI, where the sizes of datasets possessed by the parties can significantly differ. Current protocols for unbalanced PSI under the malicious security model exhibit low efficiency, rendering them impractical in real-world applications. By contrast, most efficient unbalanced PSI protocols fail to guarantee the correctness of the intersection against a malicious server and cannot even ensure the client’s privacy. The present study proposes a blockchain-based unbalanced PSI protocol with public verification and financial security that enables the client to detect malicious behavior from the server (if any) and then generate an irrefutable and publicly verifiable proof without compromising its secret. The proof can be verified through smart contracts, and some economic incentive and penalty measures are executed automatically to achieve financial security. Furthermore, we implement the proposed protocol, and experimental results demonstrate that our scheme exhibits low online communication complexity and computational overhead for the client. At the same time, the size of the generated proof and its verification complexity are both $\mathcal{O}\left(logn\right)$, enabling cost-effective validation on the blockchain. Full article

Usher syndrome (US) is a clinically and genetically heterogeneous disorder that involves three main features: sensorineural hearing loss, retinitis pigmentosa (RP), and vestibular impairment. With a prevalence of 4–17/100,000, it is the most common cause of deaf-blindness worldwide. Genetic research has provided crucial insights into the complexity of US. Among nine confirmed causative genes, MYO7A and USH2A are major players in US types 1 and 2, respectively, whereas CRLN1 is the sole confirmed gene associated with type 3. Variants in these genes also contribute to isolated forms of hearing loss and RP, indicating intersecting molecular pathways. While hearing loss can be adequately managed with hearing aids or cochlear implants (CIs), approved RP treatment modalities are lacking. Gene replacement and editing, antisense oligonucleotides, and small-molecule drugs hold promise for halting RP progression and restoring vision, enhancing patients’ quality of life. Massively parallel sequencing has identified gene variants (e.g., in PCDH15) that influence CI results. Accordingly, preoperative genetic examination appears valuable for predicting CI success. To explore genetic mutations in CI recipients and establish correlations between implant outcomes and involved genes, we comprehensively reviewed the literature to gather data covering a broad spectrum of CI outcomes across all known US-causative genes. Implant outcomes were categorized as excellent or very good, good, poor or fair, and very poor. Our review of 95 cochlear-implant patients with US, along with their CI outcomes, revealed the importance of presurgical genetic testing to elucidate potential challenges and provide tailored counseling to improve auditory outcomes. The multifaceted nature of US demands a comprehensive understanding and innovative interventions. Genetic insights drive therapeutic advancements, offering potential remedies for the retinal component of US. The synergy between genetics and therapeutics holds promise for individuals with US and may enhance their sensory experiences through customized interventions. Full article

Pectoralis Minor Syndrome (PMS) causes significant discomfort due to the compression of the neurovascular bundle within the retropectoralis minor space. Botulinum neurotoxin (BoNT) injections have emerged as a potential treatment method; however, their effectiveness depends on accurately locating the injection site. In this study, we aimed to identify optimal BoNT injection sites for PMS treatment. We used twenty-nine embalmed and eight non-embalmed human cadavers to determine the origin and intramuscular arborization of the pectoralis minor muscle (Pm) via manual dissection and Sihler’s nerve staining techniques. Our findings showed the Pm’s origin near an oblique line through the suprasternal notch, with most neural arborization within the proximal three-fourths of the Pm. Blind dye injections validated these results, effectively targeting the primary neural arborized area of the Pm at the oblique line’s intersection with the second and third ribs. We propose BoNT injections at the arborized region within the Pm’s proximal three-fourths, or the C region, for PMS treatment. These findings guide clinicians towards safer, more effective BoNT injections. Full article

This reflective case study is situated at the intersection of the literature on pedagogical partnership, child-parent research, and Critical Disability Studies. It presents a mother/daughter, faculty/student exploration of the daughter’s lived experiences of navigating, as a legally blind person, the campus and courses of a college designed for fully sighted students. After presenting our conceptual frameworks and describing, using text and a video, the daughter’s lived experience of navigating the accommodation culture on her campus, we describe the semester-long partnership process through which the video was created with the goal of moving faculty, staff, and students toward equity culture. To support others in developing such video projects on their own campuses, we draw on details of this partnership to offer guidelines for co-creating representations of the lived experiences of other students with disabilities. By synthesizing learnings from this experience and the literature noted above, we offer recommendations for transforming accommodation culture into equity culture. These recommendations include: establishing diversity as the norm in every learning context; intentionally inviting a revision of differences from deficits to resources; going beyond providing accommodations to understand students’ lived experiences; and sharing the active taking of responsibility for shifting from accommodation to equity culture. Full article

In this study, we conducted an analysis of the pedestrian safety system for crosswalks introduced in Korea to improve sustainable traffic safety. The pedestrian crossing information provision system provides information to a driver in advance when a pedestrian is detected in the driver’s blind spot when the latter is turning right at an intersection. The location analyzed was the three-way intersection in front of Yungheung Elementary School in Jeollabuk-do, and vehicle speed information for 150–160 min before and after system installation was collected. As a result of comparing and analyzing the change in the compliance rate of the spot speed and the speed limit, it was found that there was no statistical difference in the change in the spot speed, but in the absence of pedestrians, the speed increased slightly compared with that before installation. The change in the speed limit compliance rate was found to improve when pedestrian crossing information was provided. In addition, a chi-square test found that there was a difference in the speed limit compliance rate before and after system installation where pedestrians existed (when information was provided), while there was no difference in the situation where pedestrians did not exist (when information was not provided). Full article

With the development of wireless technology, signals propagating in space are easy to mix, so blind detection of communication signals has become a very practical and challenging problem. In this paper, we propose a blind detection method for broadband signals based on a weighted bi-directional feature pyramid network (BiFPN). The method can quickly perform detection and automatic modulation identification (AMC) on time-domain aliased signals in broadband data. Firstly, the method performs a time-frequency analysis on the received signals and extracts the normalized time-frequency images and the corresponding labels by short-time Fourier transform (STFT). Secondly, we build a target detection model based on YOLOv5 for time-domain mixed signals in broadband data and learn the features of the time-frequency distribution image dataset of broadband signals, which achieves the purpose of training the model. The main improvements of the algorithm are as follows: (1) a weighted bi-directional feature pyramid network is used to achieve a simple and fast multi-scale feature fusion approach to improve the detection probability; (2) the Efficient-Intersection over Union (EIOU) loss function is introduced to achieve high accuracy signal detection in a low Signal-Noise Ratio (SNR) environment. Finally, the time-frequency images are detected by an improved deep network model to complete the blind detection of time-domain mixed signals. The simulation results show that the method can effectively detect the continuous and burst signals in the broadband communication signal data and identify their modulation types. Full article