基于区域水平的指纹纹线距离估计方法

Method at Region Level for Ridge Distance Estimation in Fingerprint Images

在充分分析现有纹线距离估计方法的基础上,突破了基于整幅指纹图像或者分块图像进行纹线距离估计的传统思路,提出了一种基于区域水平的新的指纹纹线距离估计方法。首先,根据纹线方向一致性,自适应地将一幅指纹图像划分为若干个区域,提出了区域的统计窗概念。然后,在对指纹图像进行二值化后,为每个区域定义统计窗和基线,以区域为单位,采用一种距离统计方法,估计每个区域的平均纹线距离。实验证明,采用这种方法进行纹线距离估计,弱化了细节特征点和一些低质量指纹区域的存在对纹线距离估计的干扰,估计结果可靠,同时也使得该方法对图像采集分辨率具有不变性。

On the basis of analysing current methods of ridge distance estimation and not using conventional ideas to estimate ridge distance at a whole or block image, a new method at region level to estimate ridge distance of fingerprint images was put forward. A whole fingerprint image was adaptively partitioned into several regions according to the consistency of ridge directions. Based on binary valued fingerprint image, statistical window and base line of each region were defined. By using distance statistical method at the region unit, average ridge distance was estimated. Experimental results show that this method can reducede effect of reference minutiae and poor quality image on ridge distance estimation, and the ridge distance estimation results are reliable and invariable to the resolution of image acquisition system.