@article{Popoola_Lasisi_2020, title={A Biometric Fusion System of Face and Fingerprint for Enhanced Human Identification Using HOG-LBP Approach}, volume={25}, url={http://jer.unilag.edu.ng/article/view/1002}, abstractNote={<p><span class="fontstyle0">This paper presents a biometric fusion system of fingerprint and face images for Ergonomic-Based Enrolment and<br>Verification System. Features from fingerprints and faces are extracted to create a new biometric template with<br>enhanced performance and with an extra level of assurance for identification. A fusion scheme combines the<br>extracted Histogram of gradients (HOG) and local Binary Pattern (LBP) features from a subject’s fingerprint and face<br>images. Manhattan Distance is used to compare between the template in the database and the input data. The<br>difference between the database template and the input data determines the decision either to reject or accept.<br>Different "matching score thresholds" were set to evaluate the relationship between False Rejection Ratio and False<br>Acceptance Ratio which is a common measure to determine system performance level. From the experiments and<br>based on the characteristic nature of this HOG-LBP algorithm, a threshold between 75% and 80% is determined to<br>be moderate and close to the EER (Equal Error Rate) point, which is the intersection of the False Accept Rate (FAR)<br>and False Reject Rate (FRR). The system is robust enough to accommodate an increase in the threshold if a high level<br>of system confidence is required</span> </p&gt;}, number={2}, journal={Journal of Engineering Research}, author={Popoola, O.P. and Lasisi, R.A.}, year={2020}, month={Jul.}, pages={197-202} }