指纹图像质量综合评估方法

An Integrative Method for Fingerprint Image Quality Assessment

提出一种基于3个评估因子(前景比重、Gabor滤波特征值和指纹图像中心点位置)的指纹图像综合评估方法。为此,先将指纹图像分成若干ω×ω图块,计算每一图块内各像素的方向数并由此得出相关参数决定前景比重因子;再对每一图块进行8方向Gabor滤波,计算滤波结果的方差而得到Gabor滤波特征值因子;根据指纹图像相邻图块方向变化得到中心点位置因子;加权平均3个评估因子得到指纹图像质量评估结果。实验证明,该方法能准确有效地评估指纹图像的质量。

This paper proposes a new integrative method to measure fingerprint image quality on the basis of proportion of foreground, Gabor filter eigenvalue with 8 different orientations and core point position. Firstly, an image is divided into many ω×ω blocks, and several features used to determine the proportion of foreground are computed by every pixel orientation in each block; Secondly, Gabor features of each block are computed and the variance of 8 Gabor features is used to determine the Gabor filter eigenvalue; Then, core point position is determined by orientation changes between neighbouring blocks; Lastly, weighted average of above three assessment factors is used to determine fingerprint image quality. Experimental results demonstrate that the integrative method is accurate and efficient to estimate fingerprint image quality.