Altınbaş University Institutional Repository
Not a member yet
5805 research outputs found
Sort by
Tamper Detection in Multimodal Biometric Templates Using Fragile Watermarking and Artificial Intelligence
Get PDFBiometric template protection is essential for finger-based authentication systems, as template tampering and adversarial attacks threaten the security. This paper proposes a DCT-based fragile watermarking scheme incorporating AI-based tamper detection to improve the integrity and robustness of finger authentication. The system was tested against NIST SD4 and Anguli fingerprint datasets, wherein 10,000 watermarked fingerprints were employed for training. The designed approach recorded a tamper detection rate of 98.3%, performing 3–6% better than current DCT, SVD, and DWT-based watermarking approaches. The false positive rate (≤1.2%) and false negative rate (≤1.5%) were much lower compared to previous research, which maintained high reliability for template change detection. The system showed real-time performance, averaging 12–18 ms processing time per template, and is thus suitable for real-world biometric authentication scenarios. Quality analysis of fingerprints indicated that NFIQ scores were enhanced from 2.07 to 1.81, reflecting improved minutiae clarity and ridge structure preservation. The approach also exhibited strong resistance to compression and noise distortions, with the improvements in PSNR being 2 dB (JPEG compression Q = 80) and the SSIM values rising by 3%–5% under noise attacks. Comparative assessment demonstrated that training with NIST SD4 data greatly improved the ridge continuity and quality of fingerprints, resulting in better match scores (260–295) when tested against Bozorth3. Smaller batch sizes (batch = 2) also resulted in improved ridge clarity, whereas larger batch sizes (batch = 8) resulted in distortions. The DCNN-based tamper detection model supported real-time classification, which greatly minimized template exposure to adversarial attacks and synthetic fingerprint forgeries. Results demonstrate that fragile watermarking with AI indeed greatly enhances fingerprint security, providing privacy-preserving biometric authentication with high robustness, accuracy, and computational efficiency
Comparative Proteomic Analysis of Dental-Origin Stem Cells: Insights into Regenerative Potential
Get PDFTeeth are a significant source of stem cells and have clinical importance for regenerative medicine. A human tooth harbors different kinds of stem cells in the dental pulp (DPSC) or the periodontal ligament (PDLSC). Also exfoliated teeth in childhood contain a special type of stem cells in their pulp called Stem cells from Human Exfoliated Deciduous teeth (SHED). All these stem cells have features and capacities that vary depending on their niche. Here we investigated the proteomic properties of three types of stem cells that originated from human teeth. We isolated and cultured the DPSCs, PDLSCs, and SHED cells. After validating MSC populations via immunophenotyping, we performed a mass spectrometry-based proteomic approach to identify and relatively quantify whole cell and secreted proteins. Identified proteins were evaluated by using Gene Ontology and Reactome pathway analysis tools. Our data reveal that SHED cells represented inflammation, hypoxia, and nutrient deficiency-associated ontologies in both their secretome and whole-cell proteomes. The whole-cell proteome of PDLSCs consisted of differentiation and proliferation-associated molecules while their secretory molecules were mainly associated with inflammation, ECM organization, and immune response. Among dental-originated stem cells, DPSCs appeared to be the healthiest and clinically relevant in terms of proteomic properties with their proliferation, growth factor signaling, and stemness-associated molecules in their secretome and whole-cell proteome. Obtained results demonstrated that every type of stem cell from dental origin has unique proteomic features that are altered by their location and physiological conditions. The findings may help researchers improve the dental stem-cell-based regenerative medicine approaches
Boosting Multiverse Optimizer by Simulated Annealing for Dimensionality Reduction
Get PDFBackground: Because of The Multi-Verse Optimizer (MVO) has gained popularity in feature selection due to its strong global and local search capabilities. However, its effectiveness diminishes when tackling high-dimensional datasets due to the exponential growth of the search space and a tendency for premature convergence. Objective: This study aims to enhance MVO’s performance by integrating it with the Simulated Annealing Algorithm (SAA), creating a hybrid model that improves search convergence and optimizes feature selection efficiency. Methods: A High-level Relay Hybrid (HRH) architecture is proposed, where MVO identifies promising regions of the feature space and passes them to SAA for local refinement. The resulting MVOSA-FS model was evaluated on ten high-dimensional benchmark datasets from the Arizona State University (ASU) repository. Support Vector Machine (SVM) classifiers were used to assess the classification accuracy. MVOSA-FS achieved superior performance compared to six state-of-the-art feature selection algorithms: Atom Search Optimization (ASO), Equilibrium Optimizer (EO), Emperor Penguin Optimizer (EPO), Monarch Butterfly Optimization (MBO), Satin Bowerbird Optimizer (SBO), and Sine Cosine Algorithm (SCA). Results: The proposed model yielded the lowest average classification error rate (1.45%), smallest standard deviation (0.008), and most compact feature subset (0.91%). The hybrid MVOSA-FS model effectively balances exploration and exploitation, delivering robust and scalable performance in feature selection for high-dimensional data. Conclusion: This hybridization approach demonstrates improved classification accuracy and reduced computational burden
3D Printing for the Production of Food Analogues
No full textThe use of 3D printing (3DP) in food production and formulation brings many innovations to the food industry. It can be applied to many food products, such as confectionery, bakery, dairy, animal foods, and many others. With the help of food components such as carbohydrates, lipids, proteins, and hydrocolloids, different food formulations with adjusted nutritional values can be prepared uniformly for special consumer groups or advanced culinary applications and highly specific food productions. This chapter explains the current state of 3DP in food science and technology as well as the advantages and disadvantages of its usage. Current techniques used in the food industry are also elaborated. Furthermore, the recent relevant studies in the literature for different food products were critically summarized
An Efficient IoT Intrusion Detection System Based on Machine Learning Approaches
No full textArticle number : 030002
Volume editors : Albaker B.M., Ali R.M., Kwad A.M.
Conference name : International Research Conference on Engineering and Applied Sciences 2023, IRCEAS 2023
Conference code : 208810The Internet of Things (IoT) has advanced quickly and has been integrated into many different fields. With the use of this technology, gadgets have the ability for sending, receiving, and processing data automatically. IoT was rapidly accepted in many important fields since it makes life easier and boosts service quality, but privacy and security concerns are still significant problems. Intrusion Detection System (IDS) could be used as a security feature to protect IoT networks from a variety of cyber-attacks, which is a relief. This study suggests the utilization of the IDS for defending against various cyber-attacks in IoT systems. The suggested approach makes use of Random Forest (RF), Multi-layer Perceptron (MLP) to increase the detection rate, we use the pipeline to put together some processes that may be cross-validated against one another. A contemporary dataset was utilized for assessing and analyzing the performance results for validating the effectiveness of the suggested IDS approach. The evaluation findings show that the suggested IDS method may greatly increase detection performance results concerning accuracy rate while also improving detection efficiency. F1-Score, Recall, and Precision, The performance metrics demonstrate that the suggested approach produces significant outcomes, particularly when employing the pipeline with all dataset features, where the model achieved a very high result of 95.13 %
Advanced Deep Learning Models for Improved IoT Network Monitoring Using Hybrid Optimization and MCDM Techniques
Get PDFThis study addresses the challenge of optimizing deep learning models for IoT network monitoring, focusing on achieving a symmetrical balance between scalability and computational efficiency, which is essential for real-time anomaly detection in dynamic networks. We propose two novel hybrid optimization methods—Hybrid Grey Wolf Optimization with Particle Swarm Optimization (HGWOPSO) and Hybrid World Cup Optimization with Harris Hawks Optimization (HWCOAHHO)—designed to symmetrically balance global exploration and local exploitation, thereby enhancing model training and adaptation in IoT environments. These methods leverage complementary search behaviors, where symmetry between global and local search processes enhances convergence speed and detection accuracy. The proposed approaches are validated using real-world IoT datasets, demonstrating significant improvements in anomaly detection accuracy, scalability, and adaptability compared to state-of-the-art techniques. Specifically, HGWOPSO combines the symmetrical hierarchy-driven leadership of Grey Wolves with the velocity updates of Particle Swarm Optimization, while HWCOAHHO synergizes the dynamic exploration strategies of Harris Hawks with the competition-driven optimization of the World Cup algorithm, ensuring balanced search and decision-making processes. Performance evaluation using benchmark functions and real-world IoT network data highlights superior accuracy, precision, recall, and F1 score compared to traditional methods. To further enhance decision-making, a Multi-Criteria Decision-Making (MCDM) framework incorporating the Analytic Hierarchy Process (AHP) and TOPSIS is employed to symmetrically evaluate and rank the proposed methods. Results indicate that HWCOAHHO achieves the most optimal balance between accuracy and precision, followed closely by HGWOPSO, while traditional methods like FFNNs and MLPs show lower effectiveness in real-time anomaly detection. The symmetry-driven approach of these hybrid algorithms ensures robust, adaptive, and scalable monitoring solutions for IoT networks characterized by dynamic traffic patterns and evolving anomalies, thus ensuring real-time network stability and data integrity. The findings have substantial implications for smart cities, industrial automation, and healthcare IoT applications, where symmetrical optimization between detection performance and computational efficiency is crucial for ensuring optimal and reliable network monitoring. This work lays the groundwork for further research on hybrid optimization techniques and deep learning, emphasizing the role of symmetry in enhancing the efficiency and resilience of IoT network monitoring systems
Color-Spatial AutoAugment: Automated Image Augmentation and Policy Optimization
No full textConference name : 7th International Congress on Human-Computer Interaction, Optimization and Robotic Applications, ICHORA 2025
Conference city : Ankara
Conference date : 23 May 2025 - 24 May 2025
Conference code : 209351Data augmentation is a highly effective technique for improving modern image classifiers' accuracy and has continuously improved over the years. In this paper, We improve upon the previously published approach to implementing AutoAugment policies known as "Color-Spatial AutoAugment"; our implementation utilizes the best policies discovered for specific datasets. It categorizes them into color and spatial, thereby improving image classification accuracy. Applied to the CIFAR-10 dataset, our method significantly improved the performance of the networks in self-supervised training (91.13% accuracy and 0.2837 loss). We share those findings to highlight the potential of Color-Spatial AutoAugment and the improvement over similar augmentation methods
The impact of body mass index on the diagnostic and surgical outcomes in primary hyperparathyroidism
Get PDFOBJECTIVE: The aim of this study was to investigate the influence of body mass index on the diagnostic and surgical outcomes in patients undergoing parathyroidectomy for primary hyperparathyroidism. METHODS: A total of 446 patients with primary hyperparathyroidism were divided into four groups according to their body mass index: normal weight (body mass index<25 kg/m2) (n=130), overweight (25≤body mass index<30 kg/m2) (n=166), obese (30≤body mass index<35 kg/m2) (n=112), and morbidly obese (body mass index≥35 kg/m2) (n=38). Perioperative findings were compared between the groups. RESULTS: The preoperative median parathormone level in the morbidly obese group (204 pg/mL, min:max 72:1,178) was significantly lower than that in the normal-weight (246 pg/mL, min:max 60:4,262) (p=0.026) and obese (251 pg/mL, min:max 74:2,094) (p=0.012) groups. The osteoporosis rate in the normal-weight group (51%) was higher than that in the overweight (35.4%) (p=0.041) and morbidly obese (25%) (p=0.023) groups. The symptomatic hypocalcemia rate in the normal-weight group (10.2%) was significantly higher than that in the obese group (1.8%) (p=0.017). CONCLUSION: Normal-weight patients with primary hyperparathyroidism have higher blood parathormone values, higher rates of osteoporosis, and postoperative symptomatic hypocalcemia compared to patients with higher body mass index. For this reason, the surgeon should consider the possibility of symptomatic hypocalcemia after undergoing parathyroidectomy for primary hyperparathyroidism in normal-weight cases
Exploring the structural basis of crystals that affect nonlinear optical responses: An experimental and machine learning quest
No full textMachine learning can enable a computational framework to learn from data, thereby enhancing decision-making for targeted properties. Based on the significance of nonconjugated crystals as effective switches, an ML based approach has been applied to evaluate driving forces behind their polarizability/hyperpolarizability related hyper-Rayleigh Scattering (βHRS). For this, a dataset of relevant 1,3,5-triazine-2,4,6-triamine related structures in collected from peer reviewed literature to design its molecular descriptors. The designed dataset is trained on different regression models along with their cross-validation techniques include K-Fold and Leave One Group Out. It shows that Random Forest Regression can predict their polarizabilities with a fair accuracy (R2 = 0.83). Additionally, it shows its energy gaps (Egaps) ranging from 4.62 to 4.89 eV, with the smallest gap observed in ethanol. Understanding both these theoretical and experimental calculations can significantly help in selecting materials for targeted purposes, including sensors, electronic devices, and catalysis. Furthermore, insights into nucleophilic tendencies and charge distributions aids in designing new materials with tailored properties, expanding their use in various applications across chemistry, materials science, and other fields. The ML techniques prove its effectiveness to predict polarizabilities in response to its computational realm due to feature design, regression models with their cross-validations
One step synthesis of tryptophan-isatin carbon nano dots and bio-applications as multifunctional nanoplatforms
No full textThe development of natural molecule-derived carbon nano dots (CNDs) marks a significant advancement in biocompatible and sustainable nanomaterials. Tryptophan, capable of crossing the blood-brain barrier (BBB), serves as a precursor to numerous pharmacologically active compounds, while isatin and its derivatives have demonstrated anti-tumor effects, including against brain cancers. This study aimed to synthesize fluorescent CNDs from tryptophan-isatin hybrid precursor and explore their applications in glioblastoma treatment. These CNDs were characterized using techniques such as TEM, SEM-EDS, FTIR, XPS, Raman spectroscopy and UV-Vis spectrophotometry. In vitro tests using the U-87 glioblastoma cell line evaluated cell viability, affinity, and BBB permeability. The CNDs, between 4 and 7 nm in size, exhibited blue and green fluorescence, with no cytotoxic effects observed at concentrations up to 25 µg/mL. The highest BBB permeability rate was determined as 4.3 × 10⁻⁵ cm/s. Additionally, the CNDs demonstrated radiotherapeutic properties, leading to a 51 % reduction in cell viability. This research contributes to nanomedicine by introducing a novel biocompatible material with potential for targeted brain cancer imaging and therapy, while also suggesting broader applications beyond glioblastoma