Publication: FPGA implementation for finger knuckle features extraction
Loading...
Date
2023-08
Authors
FPGA implementation for finger knuckle features extraction
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
According to the global analytics software firm Fico’s identity proofing and digital banking survey, 7% of the respondents found that their identity had been stolen. Therefore, a hand-based multimodal biometric recognition plays an important role to protect their identity by using two or more biometric characteristics such as fingerprint, finger vein or finger knuckle print. Fingerprint is the most common and traditional input for recognition systems and still employed in most security devices until now. On top of that, finger knuckle print is one of the choices that can be used as input for another layer of identity recognition. Therefore, a finger knuckle features extraction and classification system is proposed in this project. To extract the features of finger knuckle print (FKP), conventional Local Binary Pattern (LBP) is used. After obtaining the LBP values for each pixel in the input image, an LBP histogram is plotted based on the frequency of occurrence of each LBP value to get the LBP feature vectors. In this project, Chi-square measure is utilized to do the classification of features. It provides a simple computation algorithm which can achieve high accuracy. The smallest Chi-square distance will be calculated and the corresponding class for test image is found. Both algorithms will be implemented into FPGA board which is DE1 SoC board. MATLAB and ModelSim software are used to verify the coding and results
obtained from FPGA board. In this project, the accuracy is calculated by using the False Acceptance Rate (FAR) and False Rejection Rate (FRR) to observe the
performance of proposed solution. It is found that the accuracy of the proposed solution is 95%.