Altınbaş University Institutional Repository
Not a member yet
5805 research outputs found
Sort by
Data and image processing for intelligent glaucoma detection and optic disc segmentation using deep convolutional neural network architecture
Get PDFArticle number : 73Glaucoma, a major reason for incurable blindness globally, is still a significant public health issue. Existing diagnostic techniques are extremely clinician-reliant and time-consuming, resulting in undue delays in detection and treatment. This work proposes a new intelligent approach using Deep Convolutional Neural Networks (DCNNs) for the detection of glaucoma and optic disc segmentation from ophthalmic medical imaging. The suggested methodology includes preprocessing retinal fundus images to improve quality, followed by feature extraction through a DCNN structure optimized for glaucoma detection. Segmentation of the optic disc is performed through the VGG-19 model. Performance metrics confirm the efficiency of the suggested approach. The constructed DCNN model proves 98.69% accuracy in differentiating between glaucomatous and non-glaucomatous eyes, far exceeding conventional methods (85–90%). The model has a 95.18% recall, which means that the majority of actual glaucoma cases are identified correctly, and an F1-score of 96.84%, which reflects a good balance between precision and recall. In addition, with an AUC-ROC of 97.63%, the model is able to distinguish glaucomatous eyes well. Experimental results validate that the proposed approach improves accuracy and efficiency compared to current methods. This work contributes to the development of automated glaucoma detection algorithms with potential clinical uses in ophthalmology and medical imaging
Efficiency of Axillary Bolster Use for Ultrasound-Guided Glenohumeral Joint Injection in MR Arthrography
No full textPurpose: The purpose of this study was to prospectively evaluate the accuracy of the ultrasonography (US)-guided posterior injection technique using an axillary bolster for magnetic resonance (MR) arthrography of the shoulder joint.
Materials and methods: This study included 60 patients (30 US-guided injections with an axillary bolster, 30 US-guided injections without an axillary bolster). There were 37 men and 23 women whose ages ranged from 17 to 64 years (mean, 36.87 years). All procedures were performed by two radiologists with less than 1 year of experience in arthrographic procedures. The accuracy of the two injection techniques was compared. Extraarticular contrast material leak was graded according to the MR arthrography findings. The number of injection attempts and the effect of contrast material extravasation rate on diagnostic quality were recorded.
Results: There were no significant differences between US-guided punctures with and without an axillary bolster in regard to pain (p = 0.39). Injections with an axillary bolster had a higher likelihood of success on the first attempt (p = 0.0031). Complete extravasation in the US-guided posterior approach technique without an axillary bolster was significantly higher than the US-guided posterior injection technique with an axillary bolster (p < 0.0001).
Conclusion: Although there is no significant difference in pain scores for both techniques, complete contrast material extravasation is seen at a higher rate in the US-guided posterior approach injection technique without the use of an axillary bolster compared to the technique used
Diş Hekimlerinin Bifosfonat Tedavisinde Farklı Dental Tedavi Yaklaşımlarına İlişkin Bilgi ve Tutumları
No full textObjective: This study evaluates oral healthcare professionals' attitudes towards managing different dental clinical scenarios considering different durations and administration routes of antiresorptive drugs. Methods: This study is a cross-sectional, Web-based survey. The first part of the survey evaluated demographic data; the second part investigated dentists' medication-related osteonecrosis of the jaw (MRONJ) treatment approach; the third part comprised clinical treatment scenario questions regarding the administration route and duration of usage. Oneway ANOVA test was used in the intergroup comparison of parameters showing normal distribution, and Tamhane's T2 test was used to determine the group that caused the difference. Student t-test was used for comparison of parameters showing normal distribution. The chi-square test, Continuity (yates) correction, and Fisher Freeman Halton test were used to evaluate qualitative data. Significance was set at P <.05. Results: 35.8% of dentists said they would not treat at-risk patients as the correct approach. More than half of the participants prefer to refer to the cases of osteonecrosis from the 1st stage to a specialist. A statistically significant difference was found between the correct answers about all five treatment scenarios for patients receiving oral bisphosphonate (BP) ≤3 years in favor of experienced dentists (P<.05). Conclusion: Within the confines of this study, dentists' knowledge of BP and osteonecrosis is moderate. Dentists with less than ten years of working experience are more cautious about patients using BP due to the risk of developing osteonecrosis when different dental treatments are performed
NATURE-AI INFUSION FOR A SUSTAINABLE IRAQ: INNOVATIVE INTERVENTIONS BEYOND 2030
No full textPURPOSE: This paper presents innovative, nature-based, and Artificial Intelligence (AI)-assisted strategies to tackle Iraq's pressing environmental challenges, particularly air pollution, climate vulnerability, and urban sustainability. DESIGN/METHODOLOGY/APPROACH: The study conceptually integrates ecological interventions, such as large-scale tree planting, AI-controlled irrigation, desert seeding, and solar energy deployment, with data-informed planning mechanisms to advance Sustainable Development Goals (SDGs) (SDG targets 11.6 and 3.9). FINDINGS: The proposed framework demonstrates how technology-led greening, educational participation, and co-ordinated policies can collectively improve environmental quality and social resilience. ORIGINALITY/VALUE: The paper introduces the Nature-AI Infusion model; this bridges environmental governance and artificial intelligence optimisation to underpin Iraq's sustainability trajectory beyond 2030. PRACTICAL IMPLICATIONS: The approach provides actionable guidance for national planners to embed AI-based monitoring and community-driven environmental initiatives within Iraq's future Voluntary National Reviews (VNRs). PAPER TYPE AND PURPOSE: This paper presents a conceptual framework with applied research insights, proposing an integrative framework that combines nature-based environmental renewal with AI-driven analytical support, guiding Iraq's transformation towards a greener, data-empowered future beyond 2030
Deep Learning-Based Bayesian Neural Network for Personalized Decision Support in Leukemia Treatment Using Medical Imaging
No full textIn the ever-evolving field of oncology, the predictive analysis of treatment outcomes has become a cornerstone for tailoring patient-specific care plans healthcare. This study introduces an innovative approach utilizing Bayesian Neural Networks (BNNs) for the prediction and management of Leukemia Treatment Using Medical Imaging. While stem cell transplantation has proven to be a viable treatment option for Leukemia, predicting post-transplant outcomes remains challenging due to the heterogeneous nature of the disease and the complex interactions of host factors, disease-specific variables, and transplant-related complications. Standard prediction models often suffer from over-fitting and inability to model complex data structures inherent in this context. Our research aims to overcome these limitations by employing a BNN, a type of neural network that provides a measure of uncertainty in predictions. The BNN model was trained on a comprehensive dataset comprising demographic, genetic, clinical, and treatment-related variables from a large cohort of Leukemia patients who underwent stem cell transplantation in Medical Imaging. The model was able to predict outcomes including survival rates, disease relapse, and graft-versus-host disease with remarkable accuracy and reliability. Importantly, the use of Bayesian approach added an extra layer of reliability to the model by providing probabilistic predictions, which are more interpretable in the clinical setting. It also offers a way to handle missing data and prevent overfitting, key challenges in healthcare datasets. This BNN-based model presents a new frontier in personalized medicine for Leukemia, empowering clinicians with precise, data-driven tools to predict treatment outcomes and customize care plans for Acute Lymphoblastic Leukemia (ALL), Chronic Lymphoblastic Leukemia (CLL), Acute Myeloid Leukemia (AML), and Chronic Myeloid Leukemia (CML) with an overall accuracy of 98.94% whereas the specificity is 9 6. 9 8 %. Further research is needed to validate the model across different population cohorts and clinical settings, and to incorporate the model into a user-friendly decision-support system for use in routine clinical practice
Permanent Magnet Synchronous Generator (BLDC) Generator Supported with Wind Turbine and Enhanced with Control Pitch Angle
No full textThe permanent magnet synchronous generator (BLDC) based on a wind energy conversion system was mathematically modeled in this work. The principle of operation of a wind energy conversion system based on a BLDC with trapezoidal EMF includes (mechanical, power electronics, and electrical parts) are described. The lightweight specific of the chosen form for the blades on the turbine has been developed and a numerical model of wind turbines has been assembled that encompasses the mechanical and lightweight processes that occur in the installation. New methods of control for the angle of incident β of a wind turbine to improve its output at low wind speeds and speed control of a BLDC engine to boost its effectiveness have been developed using PI controllers and reliable Simulink applications. Last but not least, the turbine potential efficiency can be expressed as follows: the optimum level of values of tip accelerate ratio about 8.2, the maximum torque value id determined about 0.065 and the maximum value of the estimated power coefficients (CP) for this turbine is 0.48 meaning that it works at its highest efficiency possible of 48%
Dual-Functioning Metal-Organic Frameworks: Methotrexate-Loaded Gadolinium MOFs as Drug Carriers and Radiosensitizers
No full textCancer remains a critical global health challenge, necessitating advanced drug delivery systems through innovations in materials science and nanotechnology. This study evaluates gadolinium metal-organic frameworks (Gd-MOFs) as potential drug delivery systems for anticancer therapy, particularly when combined with radiotherapy. Gd-MOFs were synthesized using terephthalic acid and gadolinium (III) chloride hexahydrate and then loaded with methotrexate (MTX). Characterization via fourier transform infrared spectroscopy (FTIR), thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), magnetic resonance imaging (MRI), and X-ray diffraction (XRD) confirmed their correct structure and stability. Effective MTX loading and controlled release were demonstrated. Anticancer effects were assessed on human healthy bronchial epithelial cells (BEAS-2B) and human lung cancer cells (A549) using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay under in vitro radiation therapy. MTX/Gd-MOF combined with radiotherapy showed a greater reduction in cancer cell viability (41.89% ± 2.75 for A549) compared to healthy cells (56.80% ± 1.97 for BEAS-2B), indicating selective cytotoxicity. These findings highlight the potential of Gd-MOFs not only as drug delivery vehicles but also as radiosensitizers, enhancing radiotherapy efficacy and offering promising evidence for their use in combinatory cancer therapies to improve treatment outcomes
Integrated System of Swarm Intelligence and Neural Network for Molecular Similarity Detection
No full textMolecular similarity, governed by the principle that “similar molecules exhibit similar properties,” is a pervasive concept in chemistry with profound implications, notably in pharmaceutical research where it informs structure-activity relationships. This study focuses on the pivotal role of molecular similarity techniques in identifying sample molecules akin to a target molecule while differing in key features. Within the realm of artificial intelligence, this paper introduces a novel hybrid system merging Swarm Intelligence (SI) behaviors (Aquila and Termites) with Neural Networks. Unlike previous applications where Aquila or Termites were used individually, this amalgamation represents a pioneering approach. The objective is to determine the most similar sample molecule in a dataset to a specific target molecule. Accuracy assessments reveal a manual evaluation accuracy of 70.58%, surging to 90% with the incorporation of Neural Networks. Additionally, a three-dimensional grid elucidates the Quantitative Structure-Activity Relationship (QSAR). The Euclidean and Manhattan Distance metrics quantify differences between molecules. This study contributes to molecular similarity assessment by presenting a hybrid approach that enhances accuracy in identifying similar molecules within complex datasets
Music Therapy may Decrease Radial Artery Spasm Rates and Increase Satisfaction during Coronary Angiography
No full textIntroduction: With the widespread use of the radial artery in catheterization procedures, radial artery spasm (RAS) is frequently considered an undesirable event. It is known that anxiety increases RAS, and listening to music helps individuals control anxiety during the procedure. This study aimed to investigate the effects of music concerts on RAS.
Methods: In this prospective study, imaging and interventional coronary catheterization procedures using the radial artery were included. One group listened to a musical recital during the procedure, while the other group was treated in a quiet environment. The demographics, procedural parameters, and complications of both groups were compared.
Results: The study included a total of 147 patients, with an average age of 51.6 ± 11.1 years. Of these, 78 patients (53%) listened to music, while 69 patients (46.9%) underwent catheterization in a quiet environment. The impact of music therapy on the RAS was found to be significant (11.5% vs. 20.3%; p=0.035). While music therapy showed a potential to reduce RAS rates, its effect was not statistically significant in multivariate analysis (p=0.055).
Conclusion: Music is a feasible, simple, and inexpensive method for reducing anxiety levels in patients. Listening to music during catheterization can reduce procedural discomfort and the frequency of undesirable events by helping people control their anxiety
A Hybrid Tree Convolutional Neural Network with Leader-Guided Spiral Optimization for Detecting Symmetric Patterns in Network Anomalies
Get PDFIn the realm of cybersecurity, detecting Distributed Denial of Service (DDoS) attacks with high accuracy is a critical task. Traditional machine learning models often fall short in handling the complexity and high dimensionality of network traffic data. This study proposes a hybrid framework leveraging symmetry in feature distribution, network behavior, and model optimization for anomaly detection. A Tree Convolutional Neural Network (Tree-CNN) captures hierarchical symmetrical dependencies, while a deep autoencoder preserves latent symmetrical structures, reducing noise for better classification. A Leader-Guided Velocity-Based Spiral Optimization Algorithm is proposed to optimize the parameters of the system and achieve better performance. A Leader-Guided Velocity-Based Spiral Optimization Algorithm is introduced to maintain a symmetrical balance between exploration and exploitation, optimizing the autoencoder, Tree-CNN, and classification thresholds. Validation using three datasets—UNSW-NB15, CIC-IDS 2017, and CIC-IDS 2018—demonstrates the framework’s superiority. The model achieves 96.02% accuracy on UNSW-NB15, 99.99% on CIC-IDS 2017, and 99.96% on CIC-IDS 2018, with near-perfect precision and recall. Despite a slightly higher computational cost, the symmetrically optimized framework ensures high efficiency and superior detection, making it ideal for real-time complex networks. These findings emphasize the critical role of symmetrical network patterns and feature selection strategies for enhancing intrusion detection performance