Palm print recognition using DWT and textural feature

This paper proposes a biometric recognition using palm print as a biometric modality, palm print consist of abandoned textural feature hence texture information are being used for recognition. The discrete wavelet transform is applied on the Region of Interest (ROI) of palm print at multi-level wavelet decomposition. Then gray level co-occurrence matrix of original as well as all sub bands of wavelet transform is formed with different orientation angle (0°, 45°, 90°, 135°), and from these matrix Harlick features are derived which used to form feature vector of original ROI. This feature vector has been used to match two images that is test image and template by using Euclidean distance. The experimentation has been carried out on publically available IIT Delhi Touch less Palm print database. The system is evaluated on the basis of performance parameter Accuracy False Acceptance Rate (FAR) and False Rejection Rate (FRR).