经验模式分解在虹膜特征提取中的应用

EMD for Iris Feature Extraction

针对Gabor滤波和小波变换在虹膜特征提取算法中参数需要优化和需要固定小波基的问题,提出了一种基于经验模式分解的虹膜特征提取算法。首先,通过对虹膜图像进行EMD分解,得到一系列固有模态函数IMF和剩余量,然后从中选择最适合特征识别的IMF分量作为特征向量。实验证明,该方法识别率高,能有效应用在虹膜识别系统中。

During the extraction of iris features by Gabor filter and the wavelet transform, Gabor filter parameters should be optimized and the wavelet basis function is fixed. In order to solve these problems, an iris feature extraction method based on EMD was proposed in this paper. First, iris signal is decomposed into several Intrinsic Mode Functions by EMD. Then IMFs of the most suitable were chosen as feature vector. Experimental results showed that the proposed method has higher recognition rate and it can be effective application of iris recognition system.