(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Embedded finger vein recognition system using Raspberry Pi

Embedded finger vein recognition system using Raspberry Pi / Leong Choon Wei
Sistem pengecaman urat jari (FVRS) mengenal pasti seseorang berdasarkan corak urat jari. Ketepatan sistem bergantung terutamanya kepada peranti perolehan imej dan ketepatan klasifikasi oleh pengelas. Berdasarkan kajian sistem yang dibangunkan sebelum ini, masalah pada peranti perolehan imej biasanya besar dalam saiz dan imej urat jari yang berkualiti rendah diperolehi. Untuk masalah dalam pengelasan, sesetengah sistem menggunakan algoritma yang sukar untuk dioptimumkan dan perlahan dalam pengiktirafan. Objektif projek ini adalah untuk membangunkan peranti dan perisian untuk FVRS. Bagi mengatasi masalah yang dikenal pasti, pencahayaan laser berhampiran inframerah dan kamera yang bersaiz kecil digunakan dalam perolehan imej, dan pengelas yang memberikan prestasi yang terbaik digunakan untuk pengenalan. Perbandingan prestasi antara beberapa pengelas dilakukan dalam MATLAB untuk mengenal pasti pengelas yang terbaik. Prosedur pemprosesan imej dibangunkan untuk menghasilkan imej urat jari yang berkualiti tinggi. Pengelas dan prosedur pemprosesan imej dilaksanakan dalam sistem menggunakan OpenCV. Berdasarkan hasil penilaian prestasi oleh sistem yang dibangunkan, ia membuktikan bahawa kaedah perolehan imej mampu untuk menangkap imej urat jari yang berkualiti baik, prosedur pemprosesan imej boleh ekstrak ROI daripada imej urat jari dan meningkatkan kejelasan corak urat, dan ketepatan identifikasi yang dicapai adalah 88%. Kesimpulannya, walaupun sistem yang dibangunkan mampu untuk menghasilkan kualiti yang tinggi imej FV tetapi ketepatan pengenalan masih disekat oleh beberapa batasan, sistem ini mempunyai potensi untuk mencapai ketepatan yang lebih tinggi dengan mengatasi batasan. _______________________________________________________________________________________________________ Finger vein recognition system (FVRS) identifies a person based on finger vein pattern. The accuracy of the system depends mainly on the image acquisition device and classification accuracy of the classifier. Based on the review of previously developed system, problems on image acquisition device usually are large in size and low quality finger vein image being acquired. For problem in classification, some systems employs algorithms that are difficult to be optimised and slow in recognition speed. The objective of this project is to develop device and software of FVRS. To overcome the identified problems, near infrared laser illumination and small size camera are used in image acquisition, and the classifier that gives the best performance is implemented for identification. The best classifier is identified by performance comparison performed in MATLAB. An image processing procedure is developed to produce high quality finger vein image. The classifier and the image processing procedure are implemented in the system using OpenCV. The performance evaluation of the developed system proves that the image acquisition method can capture good quality raw finger image, the image processing procedure can extracts region of interest (ROI) from the raw finger image with no background information and enhances the clarity of vein pattern, and the identification accuracy achieved by the developed system is up to 88%. In conclusion, although the developed system is capable to produce high quality FV image but identification accuracy is still restricted by some limitations, the system has potentials to reach higher accuracy by overcome the limitations.
Contributor(s):
Leong Choon Wei - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 875006019
Language:
English
Subject Keywords:
Finger vein recognition system (FVRS); identifies; finger vein pattern
First presented to the public:
6/1/2016
Original Publication Date:
6/14/2018
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 128
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2018-06-14 12:22:42.31
Date Last Updated
2019-01-07 11:24:32.9118
Submitter:
Mohd Jasnizam Mohd Salleh

All Versions

Thumbnail Name Version Created Date
Embedded finger vein recognition system using Raspberry Pi1 2018-06-14 12:22:42.31