Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12416/8651
Browse
23 results
Search Results
Article Self-Supervised Learning With BYOL for Non-Alcoholic Fatty Liver Disease Diagnosis Using Ultrasound Imaging(Springer London Ltd, 2025) Buktash, Ali; Gorur, Abdul KadirPurpose:The study aims to evaluate the effectiveness of Bootstrap Your Own Latent (BYOL), a self-supervised learning method for diagnosing NAFLD from ultrasound images using limited labeled data, which represents a novel approach in this domain. Self-supervised learning provides an alternative approach to traditional supervised learning by learning useful representations from unlabeled data, thereby reducing the time and cost required by radiologists to annotate images.Methods:The pre-trained ResNet-50 and ResNet-101 on the labeled ImageNet dataset were used for BYOL pre-training on ultrasound images without relying on labels. The training was conducted using default and custom augmentation, as well as balanced and imbalanced class distribution protocols. The model was then evaluated using linear and fine-tuning protocols with varying percentages of labeled data. The model was trained using three shuffled subsets, each trained 10 times. The custom augmentation set was derived by testing various augmentation settings using 100% and 1% of the labels to enhance feature learning.Results:BYOL with ResNet-101 and using the proposed custom augmentation set achieved average accuracies of 93.44%, 92.29%, and 88.49% using 100%, 10%, and 1% of the training labels across three shuffled datasets. In addition, our proposed method attained an average accuracy of 96.9% using patient-specific leave-one-out cross-validation (LOOCV).Conclusion:BYOL, with the proposed custom augmentation set, can learn effective image representations without relying on a large amount of labeled data, thereby enhancing scalability since unlabeled images are easier to acquire. It surpasses BYOL with default augmentation and training under supervised learning, especially with a low-labeled data regime.Article Enhancing Content-Based Retrieval Through an End-to-End Approach Utilizing Deep Learning and Multidimensional Indexing(Springer London Ltd, 2025) Uzel, Omer; Arslan, SerdarRecent advancements in technology, coupled with reductions in hardware and software costs, have propelled visual search applications into the spotlight, making them both popular and indispensable. Consequently, the rapid and precise retrieval of images from vast databases through image queries has become a critical task. We introduce a novel end-to-end retrieval architecture that significantly enhances retrieval performance when compared to a baseline system that conducts database searches at the video frame level. Leveraging a pre-trained convolutional neural network model, we employ unsupervised image retrieval processes to extract and store low-level features for efficient indexing. To facilitate swift and effective access, we implement a tree-based indexing structure known as VP-Tree. This structure utilizes the extracted low-level features. To make these features compatible with our system, we employ dimension reduction techniques to represent them in a lower-dimensional space. Our experiments, conducted on three benchmark datasets, demonstrate that VP-Tree consistently outperforms k-nearest neighbor (KNN) search in terms of retrieval accuracy and efficiency. Specifically, for image data set, VP-Tree achieves a precision of 56.3903, an F1-score of 68.703, and an area under the curve (AUC) of 93.518719, all slightly surpassing KNN. Similarly, for news video data set, VP-Tree attains a precision of 38.704011, an F1-score of 55.029674, and an AUC of 64.6412, again outperforming KNN. For documentary data set, VP-Tree achieves a notable improvement with a precision of 73.511723, an F1-score of 84.734013, and an AUC of 80.981328, demonstrating superior performance over KNN. In addition to accuracy, we evaluated retrieval time across different dataset sizes. While KNN performs slightly faster on smaller datasets, VP-Tree scales significantly better as dataset size increases. For 100,000 images, VP-Tree reduces retrieval time from 79.77 to 54.34 ms, and for 200,000 images, it improves performance from 108.75 to 44.63 ms, confirming its efficiency in large-scale retrieval scenarios. These results highlight VP-Tree as a robust and scalable alternative to traditional KNN-based methods, ensuring both accuracy and efficiency in large-scale image retrieval tasks.Article Low Signature UAVs: Radar Cross Section Analysis, Simulation, and Measurement in X-Band(Springer London Ltd, 2025) Unalir, Dizdar; Yalcinkaya, Bengisu; Aydin, ElifThe increasing prevalence of unmanned aerial vehicles (UAVs) is driving the development of radar systems capable of detecting them. This hampers the deployment of UAVs in military operations. While radar cross section reduction (RCSR) can be a valuable solution, the research on this subject is inadequate. This paper presents an RCSR approach adopting a shaping technique for UAVs, demonstrating the proposed approach's efficacy through simulations and actual experimental measurements performed in X-Band on a four-legged UAV model. Using electromagnetic computational instruments, the shaping is applied to the designed UAV model with parameter-based simulations, the simulated radar cross section (RCS) values are derived, and the comparative analysis of these instruments is conducted. Experimental measurements are performed in laboratory conditions using a vector network analyzer. Actual measurement results are validated by simulative findings with the examination of the influence of frequency, polarization, and aspect angle on RCS. The demonstrated measuring approach allows cost-effective and easily applicable research on RCS in X-Band, a commonly utilized frequency range in military. An average RCSR of 10 dBsm has been accomplished with the presented shaping approach.Article Citation - WoS: 18Citation - Scopus: 27Application of Bilstm-Crf Model With Different Embeddings for Product Name Extraction in Unstructured Turkish Text(Springer London Ltd, 2024) Arslan, SerdarNamed entity recognition (NER) plays a pivotal role in Natural Language Processing by identifying and classifying entities within textual data. While NER methodologies have seen significant advancements, driven by pretrained word embeddings and deep neural networks, the majority of these studies have focused on text with well-defined grammar and structure. A significant research gap exists concerning NER in informal or unstructured text, where traditional grammar rules and sentence structure are absent. This research addresses this crucial gap by focusing on the detection of product names within unstructured Turkish text. To accomplish this, we propose a deep learning-based NER model which combines a Bidirectional Long Short-Term Memory (BiLSTM) architecture with a Conditional Random Field (CRF) layer, further enhanced by FastText embeddings. To comprehensively evaluate and compare our model's performance, we explore different embedding approaches, including Word2Vec and Glove, in conjunction with the Bidirectional Long Short-Term Memory and Conditional Random Field (BiLSTM-CRF) model. Furthermore, we conduct comparisons against BERT to assess the efficacy of our approach. Our experimentation utilizes a Turkish e-commerce dataset gathered from the internet, where traditional grammatical and structural rules may not apply. The BiLSTM-CRF model with FastText embeddings achieved an F1 score value of 57.40%, a precision value of 55.78%, and a recall value of 59.12%. These results indicate promising performance in outperforming other baseline techniques. This research contributes to the field of NER by addressing the unique challenges posed by unstructured Turkish text and opens avenues for improved entity recognition in informal language settings, with potential applications across various domains.Article Citation - WoS: 40Citation - Scopus: 42Numerical and Experimental Analysis of Temperature Distribution and Melt Flow in Fiber Laser Welding of Inconel 625(Springer London Ltd, 2022) Baleanu, Dumitru; Sajadi, S. Mohammad; Ghaemi, Ferial; Fagiry, Moram A.; Tlili, Iskander; Mohammad Sajadi, S.In these days, laser is a useful and valuable tool. Low input heat, speed, accuracy, and high controllability of laser welding have led to widespread use in various industries. Nickel-based superalloys are creep-resistant materials used in high-temperature conditions. Also, these alloys have high strength, fatigue, and suitable corrosion resistance. Inconel 625 is a material that is strengthened by a complex deposition mechanism. Therefore, the parameters related to laser welding affect the microstructure and mechanical properties. Therefore, in this study, the effect of fiber laser welding parameters on temperature distribution, weld bead dimensions, melt flow velocity, and microstructure was investigated by finite volume and experimental methods. In order to detect the temperature history during continuous laser welding, two thermocouples were considered at a distance of 2 mm from the welding line. The heat energy from the laser beam was modeled as surface and volumetric heat flux. The results of numerical simulation showed that Marangoni stress and buoyancy force are the most important factors in the formation of the flow of liquid metal. Enhancing the laser power to 400 W led to the expansion of the width of the molten pool by 1.44 mm, which was in good agreement with the experimental results. Experimental results also showed that increasing the temperature from 500 degrees C around the molten pond leads to the formation of a coarse-grained austenitic structure.Article Citation - WoS: 6Citation - Scopus: 7An Island Parallel Harris Hawks Optimization Algorithm(Springer London Ltd, 2022) Dokeroglu, Tansel; Sevinc, EnderThe Harris hawk optimization (HHO) is an impressive optimization algorithm that makes use of unique mathematical approaches. This study proposes an island parallel HHO (IP-HHO) version of the algorithm for optimizing continuous multi-dimensional problems for the first time in the literature. To evaluate the performance of the IP-HHO, thirteen unimodal and multimodal benchmark problems with different dimensions (30, 100, 500, and 1000) are evaluated. The implementation of this novel algorithm took into account the investigation, exploitation, and avoidance of local optima issues effectively. Parallel computation provides a multi-swarm environment for thousands of hawks simultaneously. On all issue cases, we were able to enhance the performance of the sequential version of the HHO algorithm. As the number of processors increases, the suggested IP-HHO method enhances its performance while retaining scalability and improving its computation speed. The IP-HHO method outperforms the other state-of-the-art metaheuristic algorithms on average as the size of the dimensions grows.Article Citation - WoS: 4Citation - Scopus: 6A New Robust Harris Hawk Optimization Algorithm for Large Quadratic Assignment Problems(Springer London Ltd, 2023) Dokeroglu, Tansel; Ozdemir, Yavuz SelimHarris Hawk optimization (HHO) is a new robust metaheuristic algorithm proposed for the solution of large intractable combinatorial optimization problems. The hawks are cooperative birds and use many intelligent hunting techniques. This study proposes new HHO algorithms for solving the well-known quadratic assignment problem (QAP). Large instances of the QAP have not been solved exactly yet. We implement HHO algorithms with robust tabu search (HHO-RTS) and introduce new operators that simulate the actions of hawks. We also developed an island parallel version of the HHO-RTS algorithm using the message passing interface. We verify the performance of our proposed algorithms on the QAPLIB benchmark library. One hundred and twenty-five of 135 problems are solved optimally, and the average deviation of all the problems is observed to be 0.020%. The HHO-RTS algorithm is a robust algorithm compared to recent studies in the literature.Article Citation - WoS: 12Citation - Scopus: 15A Novel Fractional Operator Application for Neural Networks Using Proportional Caputo Derivative(Springer London Ltd, 2023) Alkan, Sertan; Baleanu, Dumitru; Altan, GokhanIn machine learning models, one of the most popular models is artificial neural networks. The activation function is one of the important parameters of neural networks. In this paper, the sigmoid function is used as an activation function with a fractional derivative approach to minimize the convergence error in backpropagation and to maximize the generalization performance of neural networks. The proportional Caputo definition is considered a fractional derivative. We evaluated three neural network models on the usage of the proportional Caputo derivative. The results show that the proportional Caputo derivative approach has higher classification accuracy than traditional derivative models in backpropagation for neural networks with and without L2 regularization.Article Citation - WoS: 1Citation - Scopus: 2Dynamical System Parameter Identification Using Deep Recurrent Cell Networks Which Gated Recurrent Unit and When(Springer London Ltd, 2021) Akagunduz, Erdem; Cifdaloz, OguzhanIn this paper, we investigate the parameter identification problem in dynamical systems through a deep learning approach. Focusing mainly on second-order, linear time-invariant dynamical systems, the topic of damping factor identification is studied. By utilizing a six-layer deep neural network with different recurrent cells, namely GRUs, LSTMs or BiLSTMs; and by feeding input/output sequence pairs captured from a dynamical system simulator, we search for an effective deep recurrent architecture in order to resolve the damping factor identification problem. Our study's results show that, although previously not utilized for this task in the literature, bidirectional gated recurrent cells (BiLSTMs) provide better parameter identification results when compared to unidirectional gated recurrent memory cells such as GRUs and LSTM. Thus, indicating that an input/output sequence pair of finite length, collected from a dynamical system and when observed anachronistically, may carry information in both time directions to predict a dynamical systems parameter.Article Citation - WoS: 7Citation - Scopus: 7Prediction of White Layer Formation in Μ-Wedm Process of Niti Shape Memory Superalloy: Fem With Experimental Verification(Springer London Ltd, 2021) Akar, Samet; Meshri, Hassan Ali M.; Seyedzavvar, Mirsadegh; Ilkhchi, Reza NajatiMicroscopic changes in the surface of nickel-titanium (nitinol) shape memory alloys (SMAs) in micro-wire electro-discharge machining (mu-WEDM) due to the formation of a resolidified layer on the machined surface, called white layer, are one of the main drawbacks in the processing of such alloys. Since these changes significantly affect the shape memory and elastic recovery characteristics of these alloys, reduction of the white layer thickness (WLT) based on the selection of optimum process parameters is essential to raise the quality of the machined parts. In this regard, a finite element model (FEM) has been developed to simulate the effects of mu-WEDM process parameters, including discharge current, pulse on-time, pulse off-time, and servo voltage, on the heat distributing in Ni55.8Ti SMA to predict the WLT. The flushing efficiency of electric discharges and the effect of flow regime of the dielectric fluid on the heat distribution in the workpiece and the formation of the WLT are analyzed. Experimental data are used to verify the accuracy of the FEM. The results show that the developed model can predict the WLT in mu-WEDM process of Ni55.8Ti SMA with an average error of 14%. The effects of discharge parameters on the formation of the WLT are discussed in details based on the results of the FEM.
- «
- 1 (current)
- 2
- 3
- »