A template selection method based on quality for fingerprint matching

Fingerprint identification system operates by acquiring a fingerprint image and comparing it with the template image, therefore the accuracy of system depends highly on the quality of template image. In this paper, we propose a method to select template based on fingerprint image quality. In the proposed method, four features firstly are extracted from a fingerprint image. And then, a support vector machine (SVM) model is trained to analyze image qualities, and lastly the fingerprint image with the best quality is chosen as the template to match for each finger. We design an experiment to compare the proposed method with two common methods in FVC2000db1, FVC2000db2 and FVC2000db3. The results of experiment show us that the proposed method is more effective to improve the identification accuracy than the other two methods.