Evaluation of a Fingerprint Identification Algorithm with SIFT Features

In fingerprint identification, the short identification time is a crucial need. The identification time can be estimated according to the number of conducted matching processes multiplied by the consumed time by a single matching process. The Matching Score Matrix is an existing identification algorithm that reduces the number of matching processes. This paper evaluates the algorithm with the features extracted from fingerprint images by SIFT. The processing time and the accuracy are compared with those of the straightforward method based on the linear search. The evaluation results prove the applicability of the algorithm to fingerprint identification using SIFT features.