图像残差的手指静脉特征间接提取的研究

Research on extracting human finger vein pattern characteristics based on residual image

根据手指静脉对近红外的吸收特性,提出了一种基于背景图像的构造以得到第一次残差图像,通过调节比例系数得到第二次残差图像,以间接的方式获取静脉的特征。经过该算法和其他直接提取算法的对比,表明该文算法具有方法简单,易于实现,特别是获取的第二次残差图像对模糊和低质量部分的图片得到了比其他算法更加合理的特征信息,真实性强,同时也达到了准确、连续、光滑的要求。

According to the absorption feature of the finger towards the near-infrared,this paper proposes an approach to obtain the first residual error image based on the construction of background image,get the second residual error image through adjusting the scale factors,and finally indirectly gain the features of the vein.This paper detailedly discusses the absorption feature of the vein towards the near-infrared and the comparison between the error-residual extraction algorithm and some other direct extraction algorithms.This method tends to be simpler and more realizable,especially in the process of obtaining the second residual error image,much more accurate feature information can be got,simultaneously conforming to the requirements of the accuracy,continuity and smoothness.