Minor finger knuckle print image edge detection using second order derivatives

Biometrics means an authentication of the person's Physiological and Behavioral Characteristics. One of the key steps in biometric based authentication is verification and identification. In physiological and behavioral based authentications are carried out using different traits such as fingerprint, finger vein, finger knuckle, iris, voice, handwriting, keystrokes, ear, EEG, ECG signals, etc. In these traits Minor finger knuckle print means backside of the finger surface and its very cheapest and easiest one for authentication system. First and second minor finger knuckle images has lot of thin lines and creases but it is not easily visible to the system for substantiation process. To overcome the above drawback image pre - processing step is very essential. Image preprocessing step segmentation is a one of the important portion. Edge detection is a pre-step of segmentation which is directly cut all the significant edges from the image correctly. In this proposed work to cut the first and second minor finger knuckle print image edges using diverse edge detection techniques. The second order derivatives are yields better results compare than first order derivatives and it's suitable for noisy Images. The output of the edge detection results are measured by various parameters such as MSE, RMSE and PSNR. According to the measurement parameters second order derivatives which give better results comparing with other datasets.