摘要： 提出一种将分级排除和综合评测相结合的多指标融合的指纹图像质量评测方法。该方法综合考虑指纹图像的局部特征和全局特征，通过从空域和频域上提取质量评测指标，采用非线性加权平均的数学模型计算指纹图像的质量得分，得到对指纹图像质量的客观评测结果。在FVC2004标准指纹库中进行实验，结果证明该方法可以准确、客观地评测指纹图像质量，评测结果与专家的视觉评测结果接近。
关键词:
指纹图像,
质量评测指标,
多指标融合,
非线性加权平均,
质量得分
Abstract: This paper presents a fingerprint image quality evaluation method based on multi-index integration, which combines progressively excluding and comprehensive evaluation. It combines local features and global features, and extracts quality evaluation indexes from spatial and frequency domain. Quality indexes are extracted and the quality scores are calculated by non-linearly summation with different weights from different quality evaluation indexes. It is tested in standard database of FVC2004, and experimental results show that the method is reasonable and effective, which is closed to experts’ evaluation.
Key words:
fingerprint image,
quality evaluation index,
multi-index integration,
non-linearly weighted average,
quality score
中图分类号:
刘莲花, 谭台哲. 多指标融合的指纹图像质量评测方法[J]. 计算机工程, 2011, 37(9): 226-228.
LIU Lian-Hua, TAN Si-Zhe. Fingerprint Image Quality Evaluation Method with Multi-index Integration[J]. Computer Engineering, 2011, 37(9): 226-228.