Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (15): 224-226.

• 工程与应用 • Previous Articles     Next Articles

Gradient based weighted averaging algorithm for compution of fingerprint directional image

FU Jia,PAN Wei,WEI Jun,HAO Chong-yang   

  1. Department of Electronic Engineering,Northwestern Polytechnical University,Xi’an 710072,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-05-21 Published:2007-05-21
  • Contact: FU Jia

一种基于加权平均梯度的指纹方向图算法

付 佳,潘 伟,位 军,郝重阳   

  1. 西北工业大学 电子信息学院,西安 710072
  • 通讯作者: 付 佳

Abstract: Directional image has important research value in a fingerprint recognition system.Conventional gradient algorithm is popular in computing of directional image but very sensitive to noise.In this paper,we propose a gradient based weighted averaging algorithm that is more resistant to noise.Because the local parallel and varying slowly features and the high correlation of neighborhoods of fingerprint,it is possible to approximate the orientation of the block using the information from its neigh-borhoods.Our experiment results suggest that the algorithm is more robust against noise in comparison with the conventional gradient algorithm.

Key words: fingerprint, directional image, weighted averaging, gradient

摘要: 方向图在指纹识别系统中具有重要的研究价值。梯度法是常见的指纹方向图求取算法,针对传统的梯度法抗噪能力差的缺点,提出了一种基于加权平均梯度的指纹方向图算法。利用指纹图像具有局部平行性和渐变性以及邻域的脊线方向相关性高的特点,通过将邻域的方向信息进行加权平均来求取块的方向。实验证明,该算法比传统的梯度法有更好的抗噪性能。

关键词: 指纹, 方向图, 加权平均, 梯度