Comparison of convolutional neural network algorithm over line tracking algorithm in finger vein recognition for improved accuracy.

The proposed study examines the efficacy of finger vein recognition for enhancing security measures across various sectors, employing a novel Convolutional Neural Network (CNN) in comparison to the Line Tracking Algorithm (LTA). This process entails data collection, training, and testing using the identified classifiers, namely Convolutional Neural Network and Line Tracking. For statistical analysis using SPSS, the outcomes of the two classifiers are grouped, each comprising 20 samples with a pre-test G-power score of 80% and a 95% confidence interval. The selected Convolutional Neural Network demonstrates enhanced digital security through finger vein recognition, achieving an accuracy of 97.4540%, whereas Line Tracking achieves an accuracy of 93.7220%. An independent sample T-test yields a p-value of p=0.002 (p<0.05), indicating statistical significance between the two groups. This underscores the superior performance of the novel Convolutional Neural Network, which outperforms the existing algorithm with an accuracy rate of 97.4540%. [ABSTRACT FROM AUTHOR]