摘要： 结合空域和频域特征的质量评估方法评估指纹图像质量。分割出指纹图像有效区域，分析该有效区域的面积、偏移度、干湿度等局部质量评估指标，并对其中心区域进行快速傅里叶变换获得全局质量评估指标，通过非线性加权融合5个质量评估指标的评估分数得到指纹图像的综合质量评估分数。实验结果表明，该方法能准确判断指纹图像的质量，有助于提高自动指纹识别系统的整体性能。

关键词: 指纹图像, 质量评估, 脊线方向, 快速傅里叶变换, 非线性融合

Abstract: This paper assesses the fingerprint image quality by combining spatial and frequency features. The valid area is separated from original fingerprint image, and the size of valid area, deflect, dry or wet extent and others other local quality indices. By Fourier spectrum analysis, a global quality index is adopted. And a weighing nonlinear fusion approach is addressed to achieve comprehensive evaluation of the final quality. Experimental results demonstrate that the algorithm can effectively evaluate the quality of fingerprint images and can significantly improve the performance of automatic fingerprint recognition system.

Key words: fingerprint image, quality evaluation, ridge line direction, Fast Fourier Transform(FFT), nonlinear fusion

中图分类号: 

• TP391.41