Special Issue Article
An Improved Principal Line-Based Alignment Refinement for Palm print Recognition
Biometric identification based on palmprint has emerged as a powerful means for recognizing a person’s identity, being used in commercial and forensic applications. Image alignment is an essential step for palmprint recognition. Most of the existing palmprint alignment algorithms make use of competitive valley detection algorithm to find some key points between fingers to establish the local coordinate system for region of interest (ROI) extraction. The ROI is subsequently used for feature extraction and matching. Currently key points based palmprint pre-processing methods provide coarse alignment only. The rotation and translation of extracted ROI image often causes the failure of genuine matching. In this paper, the palmprint verification accuracy improved by proposing an iterative closest point (ICP) algorithm to the palmprint principal lines. This compromises a more accurate alignment of palmprints by correcting efficiently the shifting, rotation and scaling variations introduced in data acquisition process. In Feature Extraction, Gabor filter is used for extracting orientation information from palmlines. The estimated parameters are then used to refine the alignment of palmprint features maps for more authentic palmprint matching. This method improves the accuracy of palmprint recognition efficiently and performs in real time environment.