Analysis Of Local Binary Pattern By Using Uniform Bins As Palm Vein Pattern Descriptor

Mohd Hayat, Nurul Atikah (2019) Analysis Of Local Binary Pattern By Using Uniform Bins As Palm Vein Pattern Descriptor. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text Analysis Of Local Binary Pattern By Using Uniform Bins As Palm Vein Pattern Descriptor - 24 Pages.pdf Download (1MB)

Abstract

Palm vein authentication technologies which reads the features of palm vein have been spread out in recent years as it offers a high accuracy identification and difficult to be forged and impersonated. However, palm vein images that were used in palm vein recognition systems were not always clear as sometimes show irregular shadings and highly saturated regions that can slow the processing time. To overcome this problem, palm vein recognition system using Uniform Local Binary Pattern (LBP) was demonstrated in this project. In this project, Spyder software with Python language and R software was implied. Python software is implemented for contrast enhancement, noise reduction and LBP implementation while R software were used for classifying palm vein pattern in K-Nearest Neighbour classifier. The sample that employed comes from two samples which are 15 sets of Chinese Academy Science Association (CASIA) and 15 sets of Rasberry Pi analysis or also called as Self dataset. The outcome were the extracted vein features from palm image and classified palm vein pattern based on subjects and their accuracy based on each dataset. For both dataset, there were two elements that were taken for the final result which were all uniform bins and selected uniform bins. The uniform bins were bin 0 to bin 25 while the selected uniform bins consist from bin 16 to bin 24. The result of the accuracy for all uniform bins and selected uniform bins from Self dataset are 87% and 53% respectively. For CASIA dataset, the percentages for all uniform bins were 60% and for selected uniform bins were 27%.

Actions (login required)

View Item