Abstract: Random sequence as the basis for information security, the quality depends on its use of random source, how to get high-quality random source research in the field of information security is one of the difficulties. Portrait facial features through the use of the difference and get (shoot) random process, a new image as a random source of portraits of random sequence generation methods; through the image acquisition process and the human biological characteristics of random noise a combination of random source. Simulation tests show that this method does not have a random sequence generated by linear correlation and nonlinear correlation, with excellent uniformity and FIPS PUB 140-2 and NIST 800-22 test pass rates to meet the needs of information security, and the method is simple and easy to implement.

Key words: Random sequence, portrait image, facial features, correlation analysis, statistical analysis

摘要： 通过利用人像面部表情特征的差异性和获取（拍摄）过程中随机性，提出了一种新的随机序列生成方法；通过将图像获取过程中的随机噪声与人体生物特征相结合的方式产生随机源。仿真测试表明，该方法产生的随机序列不具有线性相关性和非线性相关性，具有优良的均匀度和FIPS PUB 140-2及NIST 800-22测试通过率，能够满足信息安全的需要，并且方法简单,易于实现。

关键词: 随机序列, 人像图像, 面部特征, 相关性分析, 统计性分析