Fingerprint Orientation Estimation Based on Tri-Line Model

Autor: Xiaolong Zheng, Canping Zhu, Jicheng Meng
Rok vydání: 2013
Předmět:
Zdroj: Foundations and Applications of Intelligent Systems ISBN: 9783642378287
DOI: 10.1007/978-3-642-37829-4_25
Popis: As an intrinsic property of fingerprint, orientation field plays many important roles in fingerprint verification, including fingerprint image segmentation, fingerprint enhancement, and fingerprint matching. The image quality and singularity of fingerprint usually cause estimating error of orientation field. In this paper, an efficient orientation estimation method is proposed to address these problems, and we first coarsely estimate orientation field to give a rough description of fingerprint orientation, and then, we proposed a tri-line model in Fourier domain; according to the characteristic of fingerprint image patch, we accurately estimate fingerprint orientation field. The proposed approach has been tested on many public fingerprint databases (e.g., FVC2002), and the experimental results show that our method can achieve much more accuracy orientation field of fingerprints.
Databáze: OpenAIRE