Utilization of Histogram Equalization and Threshold Methods on Finger Veins

Abstract

Finger vein segmentation requires more accurate visualization to optimize patient services. This study aims to segment the finger vein features using the threshold method and compare the accuracy results. According to the experimental result, the AMT model achieves an accuracy = 57%, AGT accuracy = 62%, and the AMGT model can produce the highest accuracy at 65%. A combination of mean and Gaussian-based thresholding enhances segmentation precision and can harvest a more robust result for detecting features in complex images. The two processes between the Gaussian filter and processing using adaptive mean thresholds can produce more effective results to address finger vein segmentation issues.