Biometric Identification of EEG Signal Using Deep Learning
Subject identification in electroencephalography (EEG) plays a crucial role in various applications, such as neurorehabilitation and the Brain Computer Interface (BCI). Biometric identification is a crucial aspect of security systems, but traditional methods based on physical attributes such as...
Zapisane w:
| 1. autor: | |
|---|---|
| Format: | Praca dyplomowa |
| Wydane: |
2025
|
| Dostęp online: | https://dspace.lightring.tech/handle/123456789/20 |
| Etykiety: |
Dodaj etykietę
Nie ma etykietki, Dołącz pierwszą etykiete!
|
| _version_ | 1847021436479209472 |
|---|---|
| author | Riyadh Salam Mohammed B.Sc., |
| author_facet | Riyadh Salam Mohammed B.Sc., |
| author_sort | Riyadh Salam Mohammed B.Sc., |
| collection | DSpace |
| description | Subject identification in electroencephalography (EEG) plays a crucial role
in various applications, such as neurorehabilitation and the Brain Computer
Interface (BCI).
Biometric identification is a crucial aspect of security systems, but
traditional methods based on physical attributes such as fingerprints or facial
features have limitations and can be vulnerable to spoofing or forgery.
Therefore, there is a growing interest in exploring alternative biometric
identification techniques that offer higher security and reliability. EEG signals,
which capture the electrical activity of the brain, provide a unique and
potentially robust biometric modality for identification purposes.
Traditional EEG-based subject identification techniques frequently require
a lot of electrodes, making them cumbersome and impractical for real-world
applications. In this thesis, a new approach for subject identification in EEG
that aims to minimize the number of electrodes. A publicly available dataset
(Physionet) had been adopted and prepared by dividing the electrodes into
selected channels (16, 32odd, 32even, 64). Thus, in this work, two modules
have been proposed for personal identification (NCNN and LCNN) that were
built based on CNN architectures, exploiting the automatic feature extraction
capabilities of CNNs.
The accuracy of using 32 even channels was higher than other selection
channels when used with both models. These results clearly demonstrate the
robustness and efficacy of the proposed method for accurately identifying
individuals based on their EEG patterns. A practical and efficient solution for
subject identification in EEG is achieved by decreasing the number of
electrodes and capitalizing on the distinctive patterns captured by the odd and
even electrode groups. Based on the aforementioned system, a Graphical User
Interface (GUI) was designed as a platform and used to identify each person
separately from the corresponding EEG signal |
| format | Thesis |
| id | oai:dspace.lightring.tech:123456789-20 |
| institution | My University |
| publishDate | 2025 |
| record_format | dspace |
| spelling | oai:dspace.lightring.tech:123456789-202025-09-29T11:00:38Z Biometric Identification of EEG Signal Using Deep Learning Riyadh Salam Mohammed B.Sc., Subject identification in electroencephalography (EEG) plays a crucial role in various applications, such as neurorehabilitation and the Brain Computer Interface (BCI). Biometric identification is a crucial aspect of security systems, but traditional methods based on physical attributes such as fingerprints or facial features have limitations and can be vulnerable to spoofing or forgery. Therefore, there is a growing interest in exploring alternative biometric identification techniques that offer higher security and reliability. EEG signals, which capture the electrical activity of the brain, provide a unique and potentially robust biometric modality for identification purposes. Traditional EEG-based subject identification techniques frequently require a lot of electrodes, making them cumbersome and impractical for real-world applications. In this thesis, a new approach for subject identification in EEG that aims to minimize the number of electrodes. A publicly available dataset (Physionet) had been adopted and prepared by dividing the electrodes into selected channels (16, 32odd, 32even, 64). Thus, in this work, two modules have been proposed for personal identification (NCNN and LCNN) that were built based on CNN architectures, exploiting the automatic feature extraction capabilities of CNNs. The accuracy of using 32 even channels was higher than other selection channels when used with both models. These results clearly demonstrate the robustness and efficacy of the proposed method for accurately identifying individuals based on their EEG patterns. A practical and efficient solution for subject identification in EEG is achieved by decreasing the number of electrodes and capitalizing on the distinctive patterns captured by the odd and even electrode groups. Based on the aforementioned system, a Graphical User Interface (GUI) was designed as a platform and used to identify each person separately from the corresponding EEG signal 2025-09-29T10:51:23Z 2024 Thesis https://dspace.lightring.tech/handle/123456789/20 application/pdf |
| spellingShingle | Riyadh Salam Mohammed B.Sc., Biometric Identification of EEG Signal Using Deep Learning |
| title | Biometric Identification of EEG Signal Using Deep Learning |
| title_full | Biometric Identification of EEG Signal Using Deep Learning |
| title_fullStr | Biometric Identification of EEG Signal Using Deep Learning |
| title_full_unstemmed | Biometric Identification of EEG Signal Using Deep Learning |
| title_short | Biometric Identification of EEG Signal Using Deep Learning |
| title_sort | biometric identification of eeg signal using deep learning |
| url | https://dspace.lightring.tech/handle/123456789/20 |
| work_keys_str_mv | AT riyadhsalammohammedbsc biometricidentificationofeegsignalusingdeeplearning |