摘要
颅骨以其生理位置的隐蔽性和解剖结构的相对稳定性,近来被提出作为一种特殊的生物特征并加以识别研究。准确、高效的识别是人们追求的主要目标。基于此,首先对活体颅骨的X线照片进行二值化处理,再使用Radon变换将二维的图像转换为一维信号,最后对一维信号进行小波多分辨率分析。同时为进一步提高识别效率,提出半分处理和方向特征组方法,实现了在多尺度下准确、快速实时的身份识别。试验结果表明该方法是有效可行并适合于非活体颅骨身份鉴定和识别。
Skull identification is researched as a peculiar biometrics today for its intrinsic latent physiology position and relative stability.First,skull radiograph is processed with Radon transform following binary processing,then wavelet transform is applied to analyze the Radon transform result form multi-level.The half processing and orientation feature group is also presented to speed the recognition procedure as fast and robust.Experiment indicate that this method accommodate not only ante mortem but also postmortem.
出处
《计算机工程与应用》
CSCD
北大核心
2007年第21期242-245,共4页
Computer Engineering and Applications
