Author : S. Sri Gowthamy 1
Date of Publication :22nd March 2018
Abstract: In recent years, fingerprints are measured to be the finest method for biometric classification. These patterns are secure to use, distinctive for every person and do not change in one’s lifetime. Human fingerprints are well-off in information called minutiae, as per the individuality of the fingerprint it provides us methods of flawless detection. This work is based on the pattern type identification using features that are extracted based on minutia and are classified based on the classifiers of Neural Networks. In this paper, the images are preprocessed using thinning, termination, bifurcation, ridge to valley area process that are to obtain the final minutiae. Featured values such as mean and standard deviation are extracted from the preprocessed image. Then the classification process is done to determine which pattern is the inputted fingerprint based on their values. Classification information is essentially concerned with line patterns, whereas individual information is based on a straight or curved continuous ridgeline. The Classification processes are done using the K Nearest Neighbor Classifier using the outputted image obtained from the preprocessing and the feature extraction values. In the classification scheme, arches pattern loops pattern Whorls pattern were identified from the fingerprint images. The fingerprint classification defines the count of the pattern from the collection of images. The experimental result of the study is fully functional on the minutiae-based method and the values for identifying the different patterns
Reference :