Multi Biometric Authentication: A Secure Method with Modified Local Binary Pattern and PCA

Biometrics is an inevitable scheme in today's scenario to grant authentication for different protection systems. Still single level of biometric techniques cannot meet the expected level of accuracy, multi biometric method is much more efficient to achieve accuracy, authentication palm print technique is more easy and reliable but the requirement for perfect matching system can't be satisfied by means of it. So,by integrating the palm vein features with palm print,the perfect harmonizing can be achieved, reliable and anti spoof results for the system is possible. In this work, PCA based adaptive Gabor filter is chosen for palm print feature extraction and for vein extraction the improved version of Local Binary Pattern (LBP) is integrated with Probabilistic Neural Network (PNN). Finally the outcome of above process is logically mixed (Boolean AND) to identify the correct person.