指纹识别中动态门限值设定方法的研究

The research on dynamic threshold setting method for fingerprint identification

提出了一种在基于图像的指纹识别系统中,图像二值化中的动态门限值的确定方法。通过对几种传统门限值确定法的分析,得出传统确定门限值方法在数学计算上的弊端,进而引出基于探测概率和虚警概率确定门限值的优点,并给出了基于概率与统计学上的推导过程。这种恒虚警概率条件下的动态门限值确定法克服了传统方法的缺点,完全适应于计算机求解。因此,将该种理论与现在的处理技术和手段相结合,能够很好地满足指纹识别系统的实时性要求。

This paper puts forward a dynamic threshold determination method based on images in fingerprint identification system and image binaryzation. Through analyzing the traditional threshold determination methods, it gets the shortcomings in mathematical calcu- lation, gives advantages of determining threshold according to the detection probability and false alarm probability and presents the derivation process on the basis of probability and statistics. This dynamic threshold method under the condition of constant false alarm probability overcomes the defects of the traditional ways, and it is completely adapted to computer solving . The combination of this theory and current means and technology can meet the real-time requirements in fingerprint identification system.