Palm vein hidden under the skin and its distribution feature is hard to be stolen, which makes the palm vein recognition to be a high security biometric authentication method. Contact-less palm vein imaging can avoid ...Palm vein hidden under the skin and its distribution feature is hard to be stolen, which makes the palm vein recognition to be a high security biometric authentication method. Contact-less palm vein imaging can avoid the spread of disease, thus expanding the application range of palm vein biometric authentication devices. However, due to the different un-derstanding of the right imaging position and the change of fingers open degree, contact-less palm vein image acquisi-tion led to a certain degree of translation, rotation, scaling and shear, that is, the image deformation. Image deformation causes the imaging feature unstable. In this paper, the effect of image deformation to the stability of palm vein features is studied by some similarity parameters. First, feature points in the palm were marked, contact-less imaging and con-tact imaging of palm vein were acquired. Then, this paper calculated the similarity parameters of the contact-less imag-ing to contact imaging and gave corresponding analysis. Experimental results show that contact-less palm vein imagingwas stable, and derived the linear regression equation of relationship between sample space and the recognition rate: y =?0.000903x + 1.0332, coefficient of determination R2 = 0.9824. This research provided effective and detailed data to the study of contact-less palm vein recognition and gave powerful support to contact-less multi-feature fusion recogni-tion based on hand.展开更多
目的为了改善采集到的手掌静脉图像质量,提高静脉识别的准确性,提出一种基于Gabor滤波和Scharr边缘检测算子的手掌静脉图像增强方法。方法首先对原始手掌静脉图像截取感兴趣区域(region of interest,ROI)后进行滤波去噪,其次通过自适应...目的为了改善采集到的手掌静脉图像质量,提高静脉识别的准确性,提出一种基于Gabor滤波和Scharr边缘检测算子的手掌静脉图像增强方法。方法首先对原始手掌静脉图像截取感兴趣区域(region of interest,ROI)后进行滤波去噪,其次通过自适应直方图均衡化提高图像对比度,然后采用6个方向的Gabor滤波器提取手掌静脉纹路,并通过Scharr算子实现静脉纹路增强。采用全参考和无参考相结合的图像定量质量评价指标峰值信噪比(peak signal to noise ratio,PSNR)、均方误差(mean squared error,MSE)、灰度标准差、图像清晰度(no reference structural sharpness,NRSS)进行评估。结果本文方法PSNR为29.460,MSN为73.619,图像灰度标准差87.715、NRSS为0.341,结果表明图像失真程度最小,静脉纹路得到有效增强。结论本文方法能够有效增强手掌静脉纹路的对比度,改善图像质量。展开更多
非接触条件下单模态手部图像纹理信息有限,不利于识别。本文提出利用同一装置在多光谱下分时快速获取掌纹和掌脉图像。通过掌纹图像在静脉图像中进行掌纹主线增强处理,达到增加纹理信息、改善识别效果的目的。控制掌纹、掌脉两幅图像获...非接触条件下单模态手部图像纹理信息有限,不利于识别。本文提出利用同一装置在多光谱下分时快速获取掌纹和掌脉图像。通过掌纹图像在静脉图像中进行掌纹主线增强处理,达到增加纹理信息、改善识别效果的目的。控制掌纹、掌脉两幅图像获取间隔小于0.1s,忽略手掌细微空间移位,对获取的掌纹图像进行感兴趣区域(Region of interest,ROI)提取,并利用掌纹ROI坐标参数进行静脉图像的ROI提取。然后对两个ROI区域进行预处理,通过多层小波分解得到的掌纹高频信息快速定位掌纹主线。最后根据掌纹主线归一化后的高频分量,进行掌纹和掌脉小波系数的融合,最终得到在静脉纹理图像中增强掌纹主线纹理的融合图像。实验结果表明,融合后的图像中在保持静脉纹理信息原有状态下,掌纹主线纹理信息明显增加,识别效果得到改善。展开更多
文摘Palm vein hidden under the skin and its distribution feature is hard to be stolen, which makes the palm vein recognition to be a high security biometric authentication method. Contact-less palm vein imaging can avoid the spread of disease, thus expanding the application range of palm vein biometric authentication devices. However, due to the different un-derstanding of the right imaging position and the change of fingers open degree, contact-less palm vein image acquisi-tion led to a certain degree of translation, rotation, scaling and shear, that is, the image deformation. Image deformation causes the imaging feature unstable. In this paper, the effect of image deformation to the stability of palm vein features is studied by some similarity parameters. First, feature points in the palm were marked, contact-less imaging and con-tact imaging of palm vein were acquired. Then, this paper calculated the similarity parameters of the contact-less imag-ing to contact imaging and gave corresponding analysis. Experimental results show that contact-less palm vein imagingwas stable, and derived the linear regression equation of relationship between sample space and the recognition rate: y =?0.000903x + 1.0332, coefficient of determination R2 = 0.9824. This research provided effective and detailed data to the study of contact-less palm vein recognition and gave powerful support to contact-less multi-feature fusion recogni-tion based on hand.
文摘目的为了改善采集到的手掌静脉图像质量,提高静脉识别的准确性,提出一种基于Gabor滤波和Scharr边缘检测算子的手掌静脉图像增强方法。方法首先对原始手掌静脉图像截取感兴趣区域(region of interest,ROI)后进行滤波去噪,其次通过自适应直方图均衡化提高图像对比度,然后采用6个方向的Gabor滤波器提取手掌静脉纹路,并通过Scharr算子实现静脉纹路增强。采用全参考和无参考相结合的图像定量质量评价指标峰值信噪比(peak signal to noise ratio,PSNR)、均方误差(mean squared error,MSE)、灰度标准差、图像清晰度(no reference structural sharpness,NRSS)进行评估。结果本文方法PSNR为29.460,MSN为73.619,图像灰度标准差87.715、NRSS为0.341,结果表明图像失真程度最小,静脉纹路得到有效增强。结论本文方法能够有效增强手掌静脉纹路的对比度,改善图像质量。
文摘非接触条件下单模态手部图像纹理信息有限,不利于识别。本文提出利用同一装置在多光谱下分时快速获取掌纹和掌脉图像。通过掌纹图像在静脉图像中进行掌纹主线增强处理,达到增加纹理信息、改善识别效果的目的。控制掌纹、掌脉两幅图像获取间隔小于0.1s,忽略手掌细微空间移位,对获取的掌纹图像进行感兴趣区域(Region of interest,ROI)提取,并利用掌纹ROI坐标参数进行静脉图像的ROI提取。然后对两个ROI区域进行预处理,通过多层小波分解得到的掌纹高频信息快速定位掌纹主线。最后根据掌纹主线归一化后的高频分量,进行掌纹和掌脉小波系数的融合,最终得到在静脉纹理图像中增强掌纹主线纹理的融合图像。实验结果表明,融合后的图像中在保持静脉纹理信息原有状态下,掌纹主线纹理信息明显增加,识别效果得到改善。
