TR Dizin İndeksli Yayınlar Koleksiyonu / TR Dizin Indexed Publications Collection

Permanent URI for this collectionhttps://hdl.handle.net/20.500.14365/4

Browse

Filters

20243

Settings

Department: search.filters.department.İEÜ, Mühendislik Fakültesi, Elektrik-Elektronik Mühendisliği BölümüInstitution Author: search.filters.institutionAuthor.…

Search Results

Now showing 1 - 3 of 3
  • Article
    Citation - WoS: 5
    Citation - Scopus: 7
    Detection of Olfactory Stimulus in Electroencephalogram Signals Using Machine and Deep Learning Methods
    (Istanbul University, 2024-01-30) Akbuğday, Burak; Akbugday, S.P.; Sadikzade, R.; Akan, A.; Unal, S.; Sadighzadeh, Reza
    The investigation of olfactory stimuli has become more prominent in the context of neuromarketing research over the last couple of years. Although a few studies suggest that olfactory stimuli are linked with consumer behavior and can be observed in various ways, such as via electroencephalogram (EEG), a universal method for the detection of olfactory stimuli has not been established yet. In this study, 14-channel EEG signals acquired from participants while they were presented with 2 identical boxes, scented and unscented, were processed to extract several linear and nonlinear features. Two approaches are presented for the classification of scented and unscented cases: i) using machine learning (ML) methods utilizing extracted features; ii) using deep learning (DL) methods utilizing relative sub-band power topographic heat map images. Experimental results suggest that the olfactory stimulus can be successfully detected with up to 92% accuracy by the proposed method. Furthermore, it is shown that topographic heat maps can accurately depict the response of the brain to olfactory stimuli. © 2024 Istanbul University. All rights reserved.
  • Article
    Fluorescence Microscopy Denoizing Via Neighbor Linear Embedding
    (Istanbul University, 2024-01-31) Kırmızıay, Çağatay; Aydeniz, Burhan; Türkan, Mehmet
    One of the difficulties in studying fluorescence imaging of biological structures is the presence of noise corruption. Even though hardware- and software-related technologies have undergone continual improvement, the unavoidable effect of Poisson–Gaussian mixture type is generally encountered in fluorescence microscopy images. This noise should be mitigated to allow the extraction of valuable information from fluorescence images for various types of biological analysis. Thus, this study introduces a new and efficient learning-based denoizing approach for fluorescence microscopy. The proposed approach is based mainly on linear transformations between noise-free and noisy submanifold structures of patch spaces, benefiting from linear neighbor embeddings of local image patches. According to visual and statistical results, the developed algorithm called "neighbor linear-embedding denoizing" algorithm has a highly competitive and generally superior performance in comparison with the other algorithms used for fluorescence microscopy image denoizing in the literature. © 2024 Istanbul University. All rights reserved.
  • Article
    Citation - WoS: 1
    Detection of Alzheimer's Dementia by Using Deep Time-Frequency Feature Extraction
    (AVES, 2024-01-30) Karabiber Cura, Özlem; Türe, H. Sabiha; Akan, Aydin; Cura, Ozlem Karabiber
    Alzheimer's disease (AD), a neurological condition connected with aging, causes cognitive deterioration and has a substantial influence on a patient's daily activities. One of the most widely used clinical methods for examining how AD affects the brain is the electroencephalogram (EEG). Handcraft calculating descriptive features for machine learning algorithms requires time and frequently increases computational complexity. Deep networks provide a practical solution to feature extraction compared to handcraft feature extraction. The proposed work employs a time-frequency (TF) representation and a deep feature extraction-based approach to detect EEG segments in control subjects (CS) and AD patients. To create EEG segments'TF representations, high-resolution synchrosqueezing transform (SST) and traditional short-time Fourier transform (STFT) approaches are utilized. For deep feature extraction, SST and STFT magnitudes are used. The collected features are classified using a variety of classifiers to determine the EEG segments of CS and AD patients. In comparison to the SST method, the STFT-based deep feature extraction strategy produced improved classification accuracy between 79.56% and 92.96%.