Abstract: Biometric key solves the security problem which is not easy to remember and from external caused by the length of key. In numerous biological characteristics, enough length key can be extracted from iris feature to be used in the encryption algorithm. In allusion to the security problem existing in iris feature extraction method, a kind of key extraction method directly from iris feature is proposed. On the basis of iris preprocessing, iris feature is extracted by using Haar wavelet. The key is generated from the iris feature by using stochastic mapping function. The experimental results show that the extracted key gets through NIST random testing standard, which can meet the needs of the encryption algorithm.

Key words: iris feature, extract, stochastic mapping function, key extraction, random test

摘要： 生物特征密钥很好地解决了密钥过长不易记忆、来自于外部所带来的安全问题，而在众多生物特征中，虹膜特征能够生成足够长的密钥用于加密算法。针对现有虹膜特征密钥提取方法中密钥的安全性问题，提出一种直接从虹膜特征中提取密钥的方法。在虹膜预处理基础上，利用Haar小波提取虹膜特征；应用随机映射函数从虹膜特征中提取密钥。实验结果表明提取的密钥通过了NIST随机性测试标准，能够满足加密算法的需求。

关键词: 虹膜特征, 提取, 随机映射函数, 密钥提取, 随机测试