Bilgisayar Teknolojileri Bölümü Koleksiyonu

Permanent URI for this collectionhttps://hdl.handle.net/20.500.12514/175

Browse

Now showing 1 - 6 of 6

Citation - WoS: 17
Citation - Scopus: 19
A Class Activation Map-Based Interpretable Transfer Learning Model for Automated Detection of ADHD from fMRI Data
(Sage Journals, 2022) Uyulan, Caglar; Erguzel, Turker Tekin; Türk, Ömer; Farhad, Shams; Metin, Bariş; Tarhan, Nevzat
Automatic detection of Attention Deficit Hyperactivity Disorder (ADHD) based on the functional Magnetic Resonance Imaging (fMRI) through Deep Learning (DL) is becoming a quite useful methodology due to the curse of-dimensionality problem of the data is solved. Also, this method proposes an invasive and robust solution to the variances in data acquisition and class distribution imbalances. In this paper, a transfer learning approach, specifically ResNet-50 type pre-trained 2D-Convolutional Neural Network (CNN) was used to automatically classify ADHD and healthy children. The results demonstrated that ResNet-50 architecture with 10-k cross-validation (CV) achieves an overall classification accuracy of 93.45%. The interpretation of the results was done via the Class Activation Map (CAM) analysis which showed that children with ADHD differed from controls in a wide range of brain areas including frontal, parietal and temporal lobes.
Citation - WoS: 2
Citation - Scopus: 4
The convolutional neural network approach from electroencephalogram signals in emotional detection
(Concurrency Computation, 2021) Türk, Ömer; Özerdem, Mehmet Siraç
Although brain-computer interfaces (BCI) progress rapidly, the desired success has not been achieved yet. One of these BCI is to detect emotional states in humans. An emotional state is a brain activity consisting of hormonal and mental reasons in the face of events. Emotions can be detected by electroencephalogram (EEG) signals due to these activities. Being able to detect the emotional state from EEG signals is important in terms of both time and cost. In this study, a method is proposed for the detection of the emotional state by using EEG signals. In the proposed method, we aim to classify EEG signals without any transform (Fourier transform, wavelet transform, etc.) or feature extraction method as a pre-processing. For this purpose, convolutional neural networks (CNNs) are used as classifiers, together with SEED EEG dataset containing three different emotional (positive, negative, and neutral) states. The records used in the study were taken from 15 participants in three sessions. In the proposed method, raw channel-time EEG recordings are converted into 28 × 28 size pattern segments without pre-processing. The obtained patterns are then classified in the CNN. As a result of the classification, three emotion performance averages of all participants are found to be 88.84%. Based on the participants, the highest classification performance is 93.91%, while the lowest classification performance is 77.70%. Also, the average f-score is found to be 0.88 for positive emotion, 0.87 for negative emotion, and 0.89 for neutral emotion. Likewise, the average kappa value is 0.82 for positive emotion, 0.81 for negative emotion, and 0.83 for neutral emotion. The results of the method proposed in the study are compared with the results of similar studies in the literature. We conclude that the proposed method has an acceptable level of performance.
Citation - WoS: 8
Citation - Scopus: 14
Employing deep learning architectures for image-based automatic cataract diagnosis
(TÜBİTAK, 2021) Acar, Emrullah; Türk, Ömer; Ertuğrul, Ömer Faruk; Aldemir, Erdoğan
Various eye diseases affect the quality of human life severely and ultimately may result in complete vision loss. Ocular diseases manifest themselves through mostly visual indicators in the early or mature stages of the disease by showing abnormalities in optics disc, fovea, or other descriptive anatomical structures of the eye. Cataract is among the most harmful diseases that affects millions of people and the leading cause of public vision impairment. It shows major visual symptoms that can be employed for early detection before the hypermature stage. Automatic diagnosis systems intend to assist ophthalmological experts by mitigating the burden of manual clinical decisions and on health care utilization. In this study, a diagnosis system based on color fundus images are addressed for cataract disease. Deep learning-based models were performed for the automatic identification of cataract diseases. Two pretrained robust architectures, namely VGGNet and DenseNet, were employed to detect abnormalities in descriptive parts of the human eye. The proposed system is implemented on a wide and unique dataset that includes diverse color retinal fundus images that are acquired comparatively in low-cost and common modality, which is considered a major contribution of the study. The dataset show symptoms of cataracts in different phases and represents the characteristics of the cataract. By the proposed system, dysfunction associated with cataracts could be identified in the early stage. The achievement of the proposed system is compared to various traditional and up-to-date classification systems. The proposed system achieves 97.94% diagnosis rate for cataract disease grading.
Citation - WoS: 140
Citation - Scopus: 183
Epilepsy Detection by Using Scalogram Based Convolutional Neural Network from EEG Signals
(MDPI, 2019) Türk, Ömer; Özerdem, Mehmet Siraç
The studies implemented with Electroencephalogram (EEG) signals are progressing very rapidly and brain computer interfaces (BCI) and disease determinations are carried out at certain success rates thanks to new methods developed in this field. The effective use of these signals, especially in disease detection, is very important in terms of both time and cost. Currently, in general, EEG studies are used in addition to conventional methods as well as deep learning networks that have recently achieved great success. The most important reason for this is that in conventional methods, increasing classification accuracy is based on too many human efforts as EEG is being processed, obtaining the features is the most important step. This stage is based on both the time-consuming and the investigation of many feature methods. Therefore, there is a need for methods that do not require human effort in this area and can learn the features themselves. Based on that, two-dimensional (2D) frequency-time scalograms were obtained in this study by applying Continuous Wavelet Transform to EEG records containing five different classes. Convolutional Neural Network structure was used to learn the properties of these scalogram images and the classification performance of the structure was compared with the studies in the literature. In order to compare the performance of the proposed method, the data set of the University of Bonn was used. The data set consists of five EEG records containing healthy and epilepsy disease which are labeled as A, B, C, D, and E. In the study, A-E and B-E data sets were classified as 99.50%, A-D and B-D data sets were classified as 100% in binary classifications, A-D-E data sets were 99.00% in triple classification, A-C-D-E data sets were 90.50%, B-C-D-E data sets were 91.50% in quaternary classification, and A-B-C-D-E data sets were in the fifth class classification with an accuracy of 93.60%.
Citation - WoS: 9
Citation - Scopus: 10
FPGA simulation of chaotic tent map-based S-Box design
(Wiley Online Library, 2022) Türk, Ömer
The chaotic system has a characteristically random behavior by nature, and these systems have their own characteristics in a completely deterministic structure. This feature of a chaotic system makes it difficult to predict encryptions designed based on such a system. Thanks to this unpredictable and strong feature, maps produced from chaotic systems are an important alternative in the field of encryption. One of the structures obtained by employing chaotic maps is the substitution box. S-Box, which provides the confusion principle used in block ciphers, is the main block that dynamically replaces unencrypted data with confidential data and makes a significant contribution to ensuring high security in the encryption system. Therefore, S-Boxes hold a critical role in block ciphers. Speed and reliability are important parameters in the creation of this main block. Especially, applications performed on hardware are more reliable and high performance. Therefore, in this study, an S-Box was designed using fieldprogrammable gate arrays (FPGA) simulation from a chaotic tent map to create a fast and reliable S-Box because FPGAs offer solutions that may be important in this field considering their fast and customizable architecture. In the proposed method, the S-Box was created in 0.16 s. In addition, the dynamic properties of the chaotic tent map were analyzed with Lyapunov exponents, and the NIST SP 800-22 test was applied for the information encryption suitability of the proposed chaotic system. Also, to test the reliability of the produced S-Box structures, SAC, non-linearity, bit independence criteria, and input/output XOR distribution table metrics were implemented. The results showed that the proposed chaotic map was dynamic and passed the reliability tests successfully.
Citation - WoS: 6
Citation - Scopus: 13
Palmprint recognition system based on deep region of interest features with the aid of hybrid approach
(SpringerLink, 2023) Türk, Ömer; Çalışkan, Abidin; Acar, Emrullah; Ergen, Burhan
Palmprint recognition system is a biometric technology, which is promising to have a high precision. This system has started to attract the attention of researchers, especially with the emergence of deep learning techniques in recent years. In this study, a deep learning and machine learning-based hybrid approach has been recommended to recognize palmprint images automatically via region of interest (ROI) features. The proposed work consists of several stages, respectively. In the first stage, the raw images have been collected from the PolyU database and preprocessing operations have been implemented in order to determine ROI areas. In the second stage, deep ROI features have been extracted from the preprocessed images with the aid of deep learning technique. In the last stage, the obtained deep features have been classified by employing a hybrid deep convolutional neural network and support vector machine models. Finally, it has been observed that the overall accuracy of the proposed system has achieved very successful results as 99.72% via hybrid approach. Moreover, very low execution time has been observed for whole process of the proposed system with 0.10 s.

Browse

Browsing Bilgisayar Teknolojileri Bölümü Koleksiyonu by WoS Q "Q3"