Publication: A Palm Vein Authentication Implementation Model for Enhanced Access of Biometric Systems: A Case of Mount Kenya University Main Campus
Total Views 97
total viewsTotal Downloads 78
total downloadsDate
2023-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Mount Kenya University
Cite this Item
Abstract
One of the crucial components that contributes to the efficacy and efficiency of information systems is system integrity. One security method used to increase system integrity is biometrics. The existing fingerprint system is prone to spoofing attacks, high FRR and FAR, tear and wear of the sensor scanner. The goal of this study was to look at the integrity issues that affect the security of biometric technologies in Kenyan higher education institutions, IT security factors, implementing a new model and validating it. The implemented contactless security model sought to solve the current security problems facing the current biometric system. The study's particular goals were to look into the IT security factors that influenced biometric system integrity, review the success and failures of present biometric systems in boosting learning institution integrity, and design and validate the model. The research was guided by the extended integrated system theory which consisted of contingency and management theory. Since the contactless model had been approved by security system specialists, the researcher used an experimental and descriptive research approach. The research subject was Mount Kenya University's faculty and employees. Stratified sampling provided a true depiction of the varied population. 300 randomly chosen employees from particular departments made up the study's target population, and 169 individuals were picked for the sample using simple random sampling. The Zetech University served as the site of the pilot study. In the study area, questionnaires were used to collect the data. The researchers employed an equation for multivariate regression. Analysis of Variance (ANOVA) was used to examine the model's fitness, with a 95 percent confidence level test of significance. From the findings a strong correlation coefficient of 0.792 was obtained on objective one. This showed that the model fitted well and their statistical relationship between the variables. The correlation coefficient between the variables was at 0.792, indicating that the constructed model was more efficient in terms of data integrity. One objective two the R2 of 73.4% indicated the data fitted the model well on the assessment the IT security metrics that influenced the integrity of biometric systems in higher learning institutions since it was greater than 50%. The experiment consisted of a control group having 15 participants. From the experiment the palm vein had an FRR of 93.33% while fingerprint had 60% which demonstrated superiority in authentication accuracy. On objective three a F value of 0.714 was produced regarding the integrity of the new security model. This value is lower than the table value at (1.70) degree of freedom (10,59), which showed that there was statistical significance. 87.5% of the experts concurred that the security system satisfied the requirement for a system that can improve the integrity of the data. The researcher added feature extraction component that represented infrastructure variable in the conceptual framework. The institution should consider changing its present fingerprint security system, which failed to verify legitimate users and was therefore inconsistent with data integrity. Learning institutions should implement the contactless system that does not require physical touch to verify people, which was more useful in the current COVID 19 epidemic, which has rendered the existing fingerprint security system useless. Organizations should consider implementing live detection systems or employ cancellable biometric systems that helps overcome spoofing attacks. More research needs to be carried out on palm vein template protection in deep learning since little research has been done in the field and also new decision authentication algorithms.