Improving Accuracy of Pattern Matching Technique with Polar Harmonic Transform for Poroscopy

Abstract: Accuracy of the object classification is depend upon the feature vector of the object. Poroscopy is the method of pattern matching of the fingerprint images based upon pores. Pores of the fingerprint image are extracted by using Marker Controlled Watershed Segmentation. The feature of the pores are extracted by various methods which are rotation variant. To improve the accuracy of the pattern matching we propose an technique based upon PHT (polar harmonic transform) which is an rotation invariant transform that provide many numerical stable features. The kernel functions of PHTs consist of sinusoidal functions that are inherently computation intensive. The proposed Method will reduce the FAR and FRR.

Keywords: Poroscopy, Fingerprint images, Marker Controlled Watershed Segmentation, Polar Harmonic Transform, False Acceptance rate, False Rejection Rate.