摘要
提出了一种基于小波提升方案的虹膜识别算法。与传统基于卷积的小波变换方法相比,该方法在特征提取上计算简单,运算速度快,而且实现的是从整数到整数的变换,更有利于虹膜信息的量化。先对虹膜图像进行提升整数小波变换,然后对子图进行量化编码得到特征,采用屏蔽了噪声位的汉明距来进行模式识别。实验结果表明,基于提升整数小波变换的虹膜识别方法在识别率和等错率方面都能达到好的识别效果,在特征提取速度上也得到了较大提升,更能满足实时性的要求。
An iris recognition algorithm based on lifting integer wavelet is presented.Compared with the traditional wavelet transform which is based on convolution,the operation speed is improved and the transform from integer to integer can be realized,which is beneficial to quantify the iris information.The lifting integer wavelet transform is used to decompose the iris image and the features are extracted after the sub-image coefficients were quantified and encoded.The Hamming distance which shields noise is employed to carry on iris recognition.The experiment results show that the algorithm based on integer wavelet transform is superior on recognition rate and equal error rate,especially on the feature extraction speed.
出处
《计算机工程与应用》
CSCD
北大核心
2010年第16期207-209,共3页
Computer Engineering and Applications
关键词
虹膜识别
小波变换
提升整数小波
iris recognition
wavelet transform
lifting integer wavelet
