Altınbaş University Institutional Repository
Not a member yet
5805 research outputs found
Sort by
Improving Oil Pipeline Surveillance with a Novel 3D Drone Simulation Using Dynamically Constrained Accumulative Membership Fuzzy Logic Algorithm (DCAMFL) for Crack Detection
No full textCracks in oil pipelines pose significant risks to the environment, public safety, and the overall integrity of the infrastructure. In this paper, we propose a novel approach for crack detection in oil pipes using a combination of 3D drone simulation, convolutional neural network (CNN) feature extraction, and the dynamically constrained accumulative membership fuzzy logic algorithm (DCAMFL). The algorithm leverages the strengths of CNNs in extracting discriminative features from images and the DCAMFL’s ability to handle uncertainties and overlapping linguistic variables. We evaluated the proposed algorithm on a comprehensive dataset containing images of cracked oil pipes, achieving remarkable results. The precision, recall, and F1-score for crack detection were found to be 96.5%, 97.3%, and 95.6%, respectively. These high-performance metrics demonstrate the algorithm’s accuracy and reliability in identifying and classifying cracks. Our findings highlight the effectiveness of integrating advanced simulation techniques, deep learning, and fuzzy logic for crack detection in oil pipelines. The proposed algorithm holds promise for enhancing pipeline surveillance, improving safety measures, and extending the lifespan of oil infrastructure. Future work involves expanding the dataset, fine-tuning the CNN architecture, and validating the algorithm on large-scale pipelines to further enhance its performance and applicability
The transnational legal ordering of beneficial ownership registration
Get PDFThe lack of transparency over who owns or controls legal vehicles is a global concern. The ordering of responses to this global issue has matured over more than two decades and has reached beyond domestic to transnational legal norms. This article focuses on the process of transnational legal ordering regarding beneficial ownership registration that mandates the establishment of central government-held registries to ensure the availability of accurate and up-to-date information. This article draws on and aims to contribute to previous research into transnational legal ordering therein analytically assessing the institutionalization along the two dimensions of normative settlement and alignment. This study shows that the beneficial ownership registration transnational legal order is not yet fully institutionalized. We argue that this limitation hampers its capability to promote financial transparency and underscore the ongoing need for further advancements in transnational legal orders to combat the misuse of legal vehicles on a global scale
Flexural analysis of functionally graded sandwich plates with a novel mixed finite element formulation using quasi-3d higher order shear deformation theory
No full textThis research focuses on the modeling and analysis of Functionally Graded (FG) plates featuring a sandwich core, utilizing a generalized Higher-Order Shear Deformation Theory (HSDT) that accounts for transverse stretching effects. This framework proposed facilitates the customization of different shear functions specifically designed for functionally graded materials (FGMs), thereby ensuring the inclusion of thickness stretching considerations (epsilon z not equal 0). The governing equations are derived through the application of the Hellinger-Reissner variational principle, which guarantees the stationarity of the functional. As a result, a finite element (FE) formulation is developed, incorporating two separate field variables: displacements and stress resultants. To discretize the domain of the FG plate, four-node quadrilateral elements are utilized. Initially, the functional necessitated C1 continuity; however, by employing the two-field characteristic of the mixed finite element (MFE) approach, integration by parts is utilized, which effectively reduces the C0 continuity requirement. This approach successfully captures the nonlinear and parabolic distribution of transverse shear stresses in FG sandwich plates (FGSPs). Within the use of various shear functions, the outcomes remain theoretically aligned with elasticity-based solutions and the proposed HSDT model. Furthermore, the inclusion of transverse stretching markedly improves the accuracy of bending behavior predictions for FGSPs when compared to traditional theories
Design and optimization of 5 × 5 Network on the securing router with integrated SHA-3 & AES core in FPGA for wearable applications
Get PDFNetwork-on-Chip (NoC) has been a scalable and effective communication platform for contemporary multi-core systems, providing high-speed data exchange among IP cores with optimized power usage and security. Conventional NoC designs are plagued by latency bottlenecks and security issues, especially in wearable devices, where low power, real-time processing, and data confidentiality are essential. The current work introduces a 5 × 5 NoC router designed specifically for secure, low-power wearable systems, that employs AES-128 encryption and SHA-3 hashing to achieve end-to-end data integrity and confidentiality. Cryptographic cores are provided at the NI level based on pipeline-based encryption to minimize the processing overhead. The suggested FPGA-based NoC design reduces the utilization of logic gates by 10%, increases the speed of data processing by 5%, and decreases power consumption by 20%, which is suitable for resource-limited situations. Performance measurement under mixed traffic loads reveals that at a 1 Gbps injection rate, the system supports an end-to-end aggregate throughput of 1.05 Gbps with reduced flit transmission latency by 25% (from 120 to 90 ns) over traditional NoCs. The system also provides improved security with minimal degradation in throughput and a balance among data protection, performance, and power efficiency. These optimizations make the suggested secure NoC a strong candidate to implement real-time, low-power applications in wearable and IoT settings
Comparison of posterior support strategies with pterygoid implants for full-arch implant rehabilitation in the atrophic maxilla: a finite element study
Get PDFBackground: Rehabilitation of the atrophic maxilla is challenging due to bone resorption and anatomical limitations. Although the All-on-Four concept offers a predictable treatment strategy, posterior cantilevers may increase biomechanical risk. Pterygoid implants have been proposed as an alternative to enhance posterior support without the need for bone grafting. This study aimed to compare stress distributions in bone and prosthetic components using three different implant-supported treatment strategies for the atrophic maxilla. Methods: Three three-dimensional finite element models were developed based on cone-beam computed tomography scans of an edentulous maxilla. Model 1 included four implants following the All-on-Four protocol. Model 2 consisted of five implants, including two pterygoid implants for posterior support. Model 3 comprised six implants, also including two pterygoid implants. Stress distributions in cortical and trabecular bone, implants, abutments, and prosthetic frameworks were analyzed under oblique occlusal loading conditions. Results: Model 3 demonstrated the most favorable biomechanical profile, with reduced maximum and minimum principal stresses and strain values in both cortical and trabecular bone. Model 1 exhibited the highest stress concentrations, particularly around posterior implants (up to 110 MPa) and prosthetic components (28–45 MPa), likely due to the cantilever effect. In contrast, Model 3 showed lower maximum principal stress in the cortical bone (0.4 MPa) compared to Model 1 (1.54 MPa). Additionally, the von Mises stress in the first and second implants decreased in Model 3 (28 MPa and 63.6 MPa, respectively) compared to Model 1 (39 MPa and 110.5 MPa) and Model 2 (20 MPa and 80.0 MPa). In terms of strain distribution, all models remained within physiological thresholds but Model 3 exhibited a more balanced and uniform strain pattern, particularly around posterior implant sites. This suggests improved load transfer and reduced risk of biomechanical overload. Conclusions: Increasing the number of implants and incorporating pterygoid implants enhances biomechanical stability in the atrophic maxilla. These strategies reduce stress and strain concentrations in both bone and prosthetic components and offer a less invasive alternative to bone grafting procedures. Optimizing implant number and posterior support is critical for improving long-term success in full-arch implant rehabilitation
Development of health workers' attitude scale towards quality studies: validity and reliability study
Get PDFBackground: It is known that quality studies increase satisfaction, positively affect productivity and support corporate development. This methodological study aims to develop a scale that measures the attitudes of healthcare professionals toward quality studies. Methods: A methodological study using instrument-development and instrument verification phases. The research universe was composed of health workers working in 5 hospitals in Istanbul (N = 6308), and the sample was composed of health workers who agreed to participate in the research (n = 1013). The researchers followed the scale development stages: item pooling, expert opinion, preliminary application, validity, and reliability. Results: KMO and Bartlett’s test values of scale showed that the dataset was convenient for factor analyses (KMO = 0.976, Chi-Square = 20624.814, df = 861). In exploratory factor analysis, the 42 items comprising scale were distributed in three subscales. The confirmatory factor analysis revealed that the scale was in sufficient model fit. Cronbach’s alpha of the total scale was 0.976. Conclusion: The “Attitude of Health Workers towards Quality Studies” scale was determined to be valid and reliable. This scale could serve as a comprehensive tool for evaluating quality initiatives in healthcare. It could possess the capacity to bring together different institutions (public, private, university) in a way that fosters mutual growth. The assessment of healthcare quality initiatives could pave the way for inclusive improvements to be planned. Through all the planned enhancements, both patient and employee satisfaction can be enhanced
Fricke dosimetry analysis for cellular irradiation using 6 MV X-rays in 96-well plate format
No full textBackground: Accurate dose verification is critical in small-volume in-vitro radiobiology, where planar dosimeters do not directly report the in-well (volumetric) dose. Purpose: To assess standard Fricke solution as an in-solution dosimeter for absorbed-dose verification in 96-well plates irradiated with 6 MV photon beams. Methods: A standard Fricke solution was dispensed into 96-well plates and irradiated at 2, 4, 6, 8, and 10 Gy with a square field encompassing the plate footprint. Absorbance at 304 nm was used to compute dose via the Fricke formalism and converted to dose-to-water; results were compared against a calibrated 0.6-cc ionization chamber. Planar uniformity was checked with Gafchromic EBT-3 using triple-channel calibration with a simultaneous 0 Gy reference, and 8 × 12 well-wise dose heatmaps were generated. Results: Fricke readouts showed excellent linearity (R2 = 0.9991) and high repeatability (plate-wide CV ≤ ∼2.5 %) across 2–10 Gy, with an average difference <1 % versus the ionization chamber. Film-based verification revealed no systematic off-axis gradients; well-wise median absolute differences (Fricke vs film) were ∼5–8 % across doses. Conclusion: Fricke dosimetry provides a reliable, practical, and cost-effective in-well verification method for 96-well plate irradiations with megavoltage photons and complements planar film by directly probing the dose within the solution volume
Hybrid Ensemble Deep Learning Framework with Snake and EVO Optimization for Multiclass Classification of Alzheimer’s Disease Using MRI Neuroimaging
Get PDFAn early and precise diagnosis is essential for successful intervention in Alzheimer's disease (AD), a progressive neurological illness. In this study, we present a deep learning-based framework for multiclass classification of AD severity levels using MRI neuroimaging data. The framework integrates multiple convolutional and transformer-based architectures with a novel hybrid hyperparameter optimization strategy; Snake+EVO surpasses conventional optimizers like Genetic Algorithms and Particle Swarm Optimization by skillfully striking a balance between exploration and exploitation. A private clinical dataset yielded a classification accuracy of 99.81%for the optimized CNN model, while maintaining competitive performance on benchmark datasets such as OASIS and the Alzheimer's Disease Multiclass Dataset. Ensemble learning further enhanced robustness by leveraging complementary model strengths, and Grad-CAM visualizations provided interpretable heatmaps highlighting clinically relevant brain regions. These findings confirm that hybrid optimization combined with ensemble learning substantially improves diagnostic accuracy, efficiency, and interpretability, establishing the proposed framework as a promising AI-assisted tool for AD staging. Future work will extend this approach to multimodal neuroimaging and longitudinal modeling to better capture disease progression and support clinical translation
Ultrasound-Guided Medial Versus Lateral Approach of the Costoclavicular Block: A Prospective Randomized Study
Get PDFPurpose: Costoclavicular block is a regional anesthesia technique that involves targeting the brachial plexus in the proximal infraclavicular fossa. It can be applied via a lateral or medial approach. This randomized study aimed to compare the anesthetic efficacy and block dynamics of both approaches. Materials and Methods: Eighty-eight patients undergoing upper extremity surgery were randomly assigned to either the medial (Group M, n = 44) or lateral (Group L, n = 44) approach. The procedures were performed under ultrasound guidance. Both groups received 20 mL of local anesthetic (10 mL of 2% lidocaine and 10 mL of 0.5% bupivacaine) via an in-plane technique. In Group M, the needle was directed medially to laterally, and in Group L, it was directed laterally to medially, with the anesthetic injected at the three-cord midpoint. Block dynamics and complications were recorded. A blinded researcher assessed sensory and motor block onset every 5 minutes for 30 minutes. Surgeon-perceived anesthesia quality, additional analgesia needs, and patient satisfaction were recorded. Results: Surgical duration, procedure type, and anesthesia onset time were similar (Group M: 5.79 ± 2.14 min; Group L: 5.56 ± 1.93 min). The ultrasound scanning time was longer in Group L, but the total block performance time was not significantly different (P > 0.05). In Group M, the needle-pleura distance was greater, and the motor block onset time was shorter (P 0.05). Conclusion: The medial approach demonstrates anesthetic efficacy comparable to that of the lateral approach
Adaptive Volcano Support Vector Machine (AVSVM) for Efficient Malware Detection
Get PDFIn this paper, we propose the Adaptive Volcano Support Vector Machine (AVSVM)-a novel classification model inspired by the dynamic behavior of volcanic eruptions-for the purpose of enhancing malware detection. Unlike conventional SVMs that rely on static decision boundaries, AVSVM introduces biologically inspired mechanisms such as pressure estimation, eruption-triggered kernel perturbation, lava flow-based margin refinement, and an exponential cooling schedule. These components work synergistically to enable real-time adjustment of the decision surface, allowing the classifier to escape local optima, mitigate class overlap, and stabilize under high-dimensional, noisy, and imbalanced data conditions commonly found in malware detection tasks. Extensive experiments were conducted on the UNSW-NB15 and KDD Cup 1999 datasets, comparing AVSVM to baseline classifiers including traditional SVM, PSO-SVM, and CNN under identical computational settings. On the UNSW-NB15 dataset, AVSVM achieved an accuracy of 96.7%, recall of 95.4%, precision of 96.1%, F-1-score of 95.75%, and a false positive rate of only 3.1%, outperforming all benchmarks. Similar improvements were observed on the KDD dataset. In addition, AVSVM demonstrated smooth convergence behavior and statistically significant gains (p < 0.05) across all pairwise comparisons. These results validate the effectiveness of incorporating biologically motivated adaptivity into classical margin-based classifiers and position AVSVM as a promising tool for intelligent malware detection systems