New approach to extract palmprint lines

Biometrie authentication based on Palmprint pattern recognition has received increasing attention in recent years. The pattern of Palmprint in the hand is unique to every individual and it does not change over time. These properties of uniqueness, stability and strong immunity to forgery make it a potentially reliable biometric modality that offers greater level of security for personal authentication applications. In this work, we explored a new approach to extract the characteristics of the palm print. We used LOG-GABOR filter for lines detection. In the following for the descriptor vector dimension reduction we used the PCA technique. At the end, to evaluate the performance, we applied the Euclidean distance between the PCA vectors and a matching between the vectors that contain the descriptors significance achieved. We obtained an accuracy rate of 97.22% for the PCA-Gabor as well as an equal error rate EER = 1.5%. Validation is carried out using the data base “ CASIA multispectral palmprint”.