الفهرس 
Biometrics ApproachesOnly 14 pages are availabe for public view
 |  | from | 117 |  |  |
| | | from | 117 | | |
Abstract
Usually biometrics technology refer to those that verification / or identification of human when using behavioral like (voice, signature, Key stroke, EEG signals) or physiological features, like(fingerprint, Hand Geometry, Iris ) to solve many security problems .Despite the traditional biometric measurement that are used for identifying ,these measurement are subject to fraud or theft. For this reason, this prompted researchers to study the characteristics of an alternative biometric .These characteristics of the alternative are stronger and better in accuracy, classification, implementation time Previous studies have proven that the pattern of brain signals is unique and varies from person to person. Therefore it can use these signals as biometrics basis.
Brain signals are almost impossible to duplicate between humans . This feature helps to strengthen the security . Brain signals are one of the brain activates that have been recorded by electrode that are placed on the scalp. Electrodes are more practical comparison with second devices such as functional magnetic resonance imaging (FMRI). Usually electrodes have a primary goal of making contact and control human who have a physical disability when it use the human brain signals. Confidentiality and the difficulty in imitating it are among the most important advantages of using EEG signals .Because they represent the reflection of the individual’s mental tasks. A personal identification system based on biometric measurements depends on attempt to determine the identity of a particular person from among a closed group of persons . While verification system refer to the confirmation or denial of identity by person (one to one matching ) .These systems depend on many steps such as data / information extraction . However both modes target many application like access control systems airport checking, mobile device, computer device and laptop device .
Chapter One gives an overview on the, scope of the thesis and problem definition, related works , in addition the thesis outline . chapter Two demonstrates biometric models for human identification for example iris, finger print, voice . Also this chapter contains physiological biometrics and the types and behavioral biometrics and their types .In addition the advantages and disadvantages of each type of these measurements .
Chapter Three consist of Computational intelligence Techniques in Biometric, for example SVM, KNN,K-Means. The advantages and disadvantages of each technique were mentioned in terms of speed, accuracy, storage space. ج
Chapter Four contain Medical Aspects of Human Brain which includes the structure of the brain and brain lobes .In addition the five different frequencies(Alpha , Beta ,Gamma, Delta ,Theta ) and type of signals . Also EEG Signals acquisition and methods used to obtain signals and type of electrodes .
Chapter Five it is suggested the methods used EEG were used in our experiment for Human Identification. Dataset has been acquired from the UCI repository .Two path applied ,the first path used 13channels like (AF8, C1, C2, C3, C4, CP1, CP5, CP6, FC5,FT,P8,PO8,PZ). While second path utilized all channels with sample entropy(SaE) and Horizontal visibility graphs( HVG) methods as feature extraction . The accuracy with SaE of all channels and 13 channels are 92.6% and 83.7% respectively .While accuracy with HVG of all channels and 13 channels are 97.4% and 94.8% .respectively.
Chapter Six contain second experiment was to use the same data . Thirteen channels used such as (AF8, C1, C2, C3, C4, CP1, CP5, CP6, FC5 ,FT7, P8, PO8, PZ) when used sample entropy and graph entropy with Shannon’s entropy as feature extraction and used SVM with RBF and KNN as classifiers. The accuracy of personal identification with SVM and KNN are 90.8% and 83.7% respectively.
Chapter Seven consist of third experiment, (EEG)data have been used to human identification by computing sample entropy and graph entropy as feature extractions. Used two classifier types which are K-Nearest Neighbors (K-NN) and Support Vector Machine (SVM). Python and MATLAB software were used in this study and EEG data was collected by UCI repository .MATLAB used when Thirteen channels was applied as feature extraction . The experimental results show that, Python software
accuracy of KNN and SVM, were 85.2% and 91.5% respectively.
While chapter eight present result summary, conclusion remarks and future work.