Log-gabor filter based finger vein biometric system using modified repeated line tracking algorithm
Loading...
Date
2018-07-01
Authors
Amir Hajian
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The performance of finger vein recognition system relies on the quality of captured image. Although the classical linear Un-sharp mask can enhance the dark and shadowy parts of finger vein image, but the enhanced image suffers two drawbacks. First, the halo effects that appears around sharper areas of image. Second, the noises which exist in image are over enhanced. This study modifies the classical linear Un-sharp mask with use of Log-Gabor filter. This Modified Un-sharp Mask (MUM) enhances the contrast and sharpness of image without aforementioned drawbacks. This study, introduced a pre-processing stage in the finger vein verification system which first, applies Contrast Limit Adaptive Histogram Equalization (CLAHE) method on input image then use MUM technique in order to enhance the sharpness and contrast of finger vein image. The results of extracted feature show the excellent improvement in detection of vein details by using the proposed pre-processing method. The Modified Repeated Line Tracking (MRLT) is used as feature extraction method and Support Vector Machine (SVM) is used as classifier. The Equal Error Rate (EER) is used as performance evaluation in this study. The EERs for the verification system at three training data is observed to be 16.66% for original image, 14.22% for CLAHE enhanced image and 6.28% for proposed method (CLAHE then MUM).