Accurate Biometric Palm Print Recognition Using ResNet50 algorithm Over X Gradient Boosting Algorithm

Abstract

The aim of this research is to enhance the accuracy of biometric palm print identification by using the Novel ResNet50 Algorithm as compared to the X Gradient Boosting. Materials and Methods: In this study, the ResNet50 and X Gradient Boosting algorithms were compared using a sample size of 10 for each algorithm, resulting in a total sample size of 20. The comparison was carried out with a G Power of 0.8 and a confidence interval (CI) of 95% to ensure statistical significance. For this study the Birjand University Mobile Palmprint Database (BMPD) dataset was collected from the Kaggle repository, which includes a total of 1640 images containing both left and right-hand palmprints. Result: According to the results, the ResNet50 algorithm achieved a higher accuracy rate (94.7%) compared to the X Gradient Boosting algorithm (92.4%) in identifying and measuring the images. The statistical analysis indicated a significant difference between the Novel ResNet50 algorithm and X Gradient Boosting, with a pvalue of 0.003 (Independent sample T-test p<0.05). This suggests that the ResNet50 algorithm outperformed the X Gradient Boosting algorithm in this experiment. According to the study’s findings, ResNet50 is more effective in accurately identifying biometric palm prints compared to X Gradient Boosting.

Palmprint Identification Using Dolphin Optimization

“…In 2023 Kumar and Kumar proposed a method for palm recognition using two classification algorithms Novel ResNet50 and X Gradient Boosting, where a comparison was made between the algorithms ResNet50 and X Gradient Boosting using a sample size of 10 per algorithm, which led to a total sample size of 20, and according to the results of the study, ResNet50 is more effective in accurately identifying biometric palm fingerprints compared to the Gradient Boosting X algorithm. The statistical analysis of this study was carried out using IBM SPSS, this proposed system was tested on the database of palm print (BMPD) collected from the kaggle warehouse, which includes a total of 1640 images containing the fingerprints of the right and left hand, and according to the results, the ResNet50 algorithm achieved the highest rating accuracy rate of (94.7%) compared to the X Gradient Boosting algorithm, in which the classification accuracy reached (92.4%) [22].…”