Biometric Authentication of Person using finger knuckle

Finger knuckle print can be extremely beneficial for person identity. Finger knuckle as a new biometric approach which supplying huge scope for researchers in few years. In this system we present a finger knuckle authentication method by using subspace techniques. In implemented system we use the three techniques with one classifier, first we use Gabor filter in preprocessing step for removing the noise from acquired image and we get the noise free image. Secondly we use PCA for feature extraction and then last we use the LDA and PNN classifier for matching purpose. Result obtained from PNN classifier which gives the approximately 90% recognition rate. Comparatively it improves the efficiency and accuracy of the Finger Knuckle recognition.