Yazar "Cizmeci, Huseyin" seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • Küçük Resim Yok
    Öğe
    Channel selection and feature extraction on deep EEG classification using metaheuristic and Welch PSD
    (Springer, 2022) Cizmeci, Huseyin; Ozcan, Caner; Durgut, Rafet
    Brain computer interfaces are important for different application domain such as medical, natural interfaces and entertainment. Besides the difficulty of gathering data from the human brain via different channel probs, preprocessing of data is another different and important task that must be solved in order to get better achievement. Selection of the most active channels is an important problem to achieve high classification accuracy. Metaheuristics are good solutions for selecting the optimal subset from the original set, as they have the ability to obtain an acceptable solution in a reasonable time. At the same time, it is necessary to use the correct feature extraction method so that the data can be properly represented. In addition, traditional deep learning methods used for emotion recognition ignore the spatial properties of EEG signals. This reduces the classification accuracy. In this study, we used artificial bee colony optimization algorithm on the seed dataset to increase the classification accuracy. We implemented and tested four different variations of this algorithm. Then, we extracted the features of the obtained channels with the Welch PSD method. We used enhanced capsule network as a machine learning algorithm and showed the best configuration to solve the problem. At the end of the process, 99.98% training and 99.83% test accuracy rates were obtained.
  • Küçük Resim Yok
    Öğe
    Enhanced deep capsule network for EEG-based emotion recognition
    (Springer London Ltd, 2023) Cizmeci, Huseyin; Ozcan, Caner
    Recently, it has become very popular to use electroencephalogram (EEG) signals in emotion recognition studies. But, EEG signals are much more complex than image and audio signals. There may be inconsistencies even in signals recorded from the same person. Therefore, EEG signals obtained from the human brain must be analyzed and processed accurately and consistently. In addition, traditional algorithms used to classify emotion ignore the neighborhood relationship and hierarchical order within the EEG signals. In this paper, a method including selection of suitable channels from EEG data, feature extraction by Welch power spectral density estimation of selected channels and enhanced capsule network-based classification model is presented. The most important innovation of the method is to adjust the architecture of the capsule network to adapt to the EEG signals. Thanks to the proposed method, 99.51% training and 98.21% test accuracy on positive, negative and neutral emotions were achieved in the Seed EEG dataset. The obtained results were also compared and evaluated with other state-of-the-art methods. Finally, the method was tested with Dreamer and Deap EEG datasets.