一种基于Morlet小波变换的虹膜识别算法

Recognition of Iris Using the Morlet Wavelet Transform

虹膜识别包括虹膜定位、特征提取以及模式匹配几个步骤。文章提出了基于虹膜灰度梯度分析的新定位算法和基于Morlet小波变换的特征提取算法。首先对沿瞳孔半径方向展开的虹膜图像通过寻找灰度梯度最大值位置的方法进行虹膜定位;然后根据虹膜生理的特点对虹膜图像进行分区,对不同的虹膜区域采用一维和二维Morlet复小波变换相结合的特征提取算法,并用二比特格雷编码来表征提取的虹膜纹理的相位信息;最后通过计算虹膜间的Hamming距进行匹配,最终实现虹膜识别。实验结果表明,与现有算法相比,该算法识别速度快,提取特征的效果好,在实验室身份认证系统中表现出很好的识别效果。

Iris recognition comprises some kernel procession as iris localization,acquisition of texture character,coding and recognition.There are some new ideas about the iris localization and the character extract using the Morlet wavelet transform in this paper.Firstly,this paper promotes a new algorithm on the basis of previous two-step iris location combining coarse location with fine location by analyzing of the change of gradation of grey.Secondarily,algorithms attain the phase information of the texture character points with one dimension and two dimension Morlet complex wavelet transform according to the different iris area.At last,by calculating the hamming distance between two compared irises we give a judgment.The results show that this approach has a very high speed of operation and good result and can be used to actually iris recognition system.