Development of biosignals-based multimodal biometric system
The success of any biometric system is dependent on the pre-processing, features and classification algorithms exploits. There is a need for increase in reliability and security motivated by the fact that there is no unique technology that can be applied for all possible scenarios. In this thesis, t...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2014
|
| Online Access: | https://eprints.ums.edu.my/id/eprint/38181/1/24%20PAGES.pdf https://eprints.ums.edu.my/id/eprint/38181/2/FULLTEXT.pdf |
| _version_ | 1825715490921644032 |
|---|---|
| author | Osamah Sadeq Alhamdani |
| author_facet | Osamah Sadeq Alhamdani |
| author_sort | Osamah Sadeq Alhamdani |
| collection | UMS |
| description | The success of any biometric system is dependent on the pre-processing, features and classification algorithms exploits. There is a need for increase in reliability and security motivated by the fact that there is no unique technology that can be applied for all possible scenarios. In this thesis, the work is focused on the use of the heart which can reproduce two biological signals, namely electrocardiogram (ECG) and Phonocardiogram (PCG). In addition, the traditional unimodal biometric system called Speaker Recognition is also implemented for the performance comparison with the previous models. Several problems make speakers identification by their voice difficult and these include isolated, connected or continuous speech, task and model constrain, vocabulary size, speaker dependent or independent, variability of speech and environmental noise. In one part of the thesis the usage of heart sounds as physiological tract was analyzed with multimodal and fusion techniques to compensate the above weakness. This thesis also implements a new technique based on the fiducial points of the ECG, which results in significant improvement performance for biometric matching. Experimental results show that the recognition rate for the Heart Sound Identification (HSI) model is 81.9%, while the rate for the Speaker Identification (SI) model is 99.3% for 20 clients. Heart Sound Verification (HSV) provides an average equal error rate (EER) of 13.8% while the average EER for the Speaker Verification model (SV) is 2.1 %. ECG Identification (ECGI) on the other hand provides an accuracy of 98. 7% and ECG Verification (ECGV) EER of 4.2%. In order to reach a higher security level, an alternative multimodal and a fusion technique is implemented into the system. Through the performance analysis of the three biometric systems and their combination using two multimodal biometric score level fusion, this thesis finds the optimal combination of those systems. The best performance of the work is based on simple-sum score fusion with a piecewiselinear normalization technique which provides an EER of 0. 7%. |
| first_indexed | 2024-03-06T03:27:37Z |
| format | Thesis |
| id | ums.eprints-38181 |
| institution | Universiti Malaysia Sabah |
| language | English English |
| last_indexed | 2024-03-06T03:27:37Z |
| publishDate | 2014 |
| record_format | dspace |
| spelling | ums.eprints-381812024-02-09T03:10:42Z https://eprints.ums.edu.my/id/eprint/38181/ Development of biosignals-based multimodal biometric system Osamah Sadeq Alhamdani The success of any biometric system is dependent on the pre-processing, features and classification algorithms exploits. There is a need for increase in reliability and security motivated by the fact that there is no unique technology that can be applied for all possible scenarios. In this thesis, the work is focused on the use of the heart which can reproduce two biological signals, namely electrocardiogram (ECG) and Phonocardiogram (PCG). In addition, the traditional unimodal biometric system called Speaker Recognition is also implemented for the performance comparison with the previous models. Several problems make speakers identification by their voice difficult and these include isolated, connected or continuous speech, task and model constrain, vocabulary size, speaker dependent or independent, variability of speech and environmental noise. In one part of the thesis the usage of heart sounds as physiological tract was analyzed with multimodal and fusion techniques to compensate the above weakness. This thesis also implements a new technique based on the fiducial points of the ECG, which results in significant improvement performance for biometric matching. Experimental results show that the recognition rate for the Heart Sound Identification (HSI) model is 81.9%, while the rate for the Speaker Identification (SI) model is 99.3% for 20 clients. Heart Sound Verification (HSV) provides an average equal error rate (EER) of 13.8% while the average EER for the Speaker Verification model (SV) is 2.1 %. ECG Identification (ECGI) on the other hand provides an accuracy of 98. 7% and ECG Verification (ECGV) EER of 4.2%. In order to reach a higher security level, an alternative multimodal and a fusion technique is implemented into the system. Through the performance analysis of the three biometric systems and their combination using two multimodal biometric score level fusion, this thesis finds the optimal combination of those systems. The best performance of the work is based on simple-sum score fusion with a piecewiselinear normalization technique which provides an EER of 0. 7%. 2014 Thesis NonPeerReviewed text en https://eprints.ums.edu.my/id/eprint/38181/1/24%20PAGES.pdf text en https://eprints.ums.edu.my/id/eprint/38181/2/FULLTEXT.pdf Osamah Sadeq Alhamdani (2014) Development of biosignals-based multimodal biometric system. Masters thesis, Universiti Malaysia Sabah. |
| spellingShingle | Osamah Sadeq Alhamdani Development of biosignals-based multimodal biometric system |
| title | Development of biosignals-based multimodal biometric system |
| title_full | Development of biosignals-based multimodal biometric system |
| title_fullStr | Development of biosignals-based multimodal biometric system |
| title_full_unstemmed | Development of biosignals-based multimodal biometric system |
| title_short | Development of biosignals-based multimodal biometric system |
| title_sort | development of biosignals based multimodal biometric system |
| url | https://eprints.ums.edu.my/id/eprint/38181/1/24%20PAGES.pdf https://eprints.ums.edu.my/id/eprint/38181/2/FULLTEXT.pdf |
| work_keys_str_mv | AT osamahsadeqalhamdani developmentofbiosignalsbasedmultimodalbiometricsystem |