基于小波分析的颅骨复合特征计算与识别

Wavelet-based Skull Contour Identification

颅骨作为一个新的生物特征,其突出优点是不受表情的干扰和光线的约束。我们按照快速准确性要求,提出一种多尺度复合特征识别方法。首先对比颅骨最大夹角,其次进行边缘的多尺度分析,最后以颅腔面积作为辅助识别。试验结果证明,该方法能快速有效的实现活体或非活体的颅骨特征计算和识别。

Biometric recognition methods for human identification have received researchers＇ attention. Seeing that the prominent characteristics of skull biometrics are not affected by facial expression and illumination, we present a new approach compounded of several features. Firstly, comparing the crania included angle; secondly, sampling the edge as the skull feature and analyzing the skull feature in multi-resolution; lastly, calculating the crania cavity area and regularizing the recognition rules. This method supports human identification in cases not only ante mortem but also postmortem.