Search In this Thesis
   Search In this Thesis  
العنوان
Using finger vein to improve personal identification /
المؤلف
Al-Juboria, Marwan Ali Jumaah.
هيئة الاعداد
باحث / مروان على جمعة البجورى
مشرف / مجدى زكريا رشاد
مشرف / شريف السيد حسين
الموضوع
Biometric identification. Identification. Fingerprints.
تاريخ النشر
2016.
عدد الصفحات
78 p. :
اللغة
الإنجليزية
الدرجة
ماجستير
التخصص
Computer Science (miscellaneous)
تاريخ الإجازة
1/1/2016
مكان الإجازة
جامعة المنصورة - كلية الحاسبات والمعلومات - Computer sciences
الفهرس
Only 14 pages are availabe for public view

from 78

from 78

Abstract

Recognition systems become more complicated over time. These systems are derived from features which can be extracted from different body members using extractor methods. Finger vein is suitable member that could be used to violate the weakness of finger print. Conventional extractor methods like matched filter and morphological methods can extricate patterns if the widths of veins are steady whereas repeated line tracking method extract vein patterns from a hazy picture. These strategies can’t remove veins that are smaller extensive than the accepted widths which corrupts the precision of the individual recognizable proof or can’t adequately extricate flimsy veins on the grounds. In turn, they have proposed a system that tackles these issues by checking the shape of the picture profiles and stressing just the centerlines of veins. Our system for distinguishing the most extreme bend positions is hearty against transient vacillations in vein width and splendor. This thesis introduces a finger vein recognition model based on Histogram of Gradient (HOG) and multi class Support Vector Machine (SVM) with accuracy 95% and finger vein recognition is powered by using Gabor filter with classifier powered by multi class support vector machine with accuracy 96.67%. The proposed method has great enhancement impact over relative to accuracy, sensitivity, F-measure and precision during evaluation.