基于局部灰度极小值的指静脉图像分割方法

A Segmentation Method for Finger Vein Image Based on Local Gray Minimum

为了解决在光照不均匀、对比度低和指节纹干扰等情况下存在的手指静脉纹线分割效果不好的问题,文中提出一种基于局部灰度极小值的指静脉检测方法。根据指静脉纹线的走向选取垂直于指静脉方向的模板,该检测模板由三个子模板组成。由于静脉处较其周围邻域的灰度值较低,当检测模板由上至下逐点检测时,中间子模板的灰度值之和小于其他两个子模板的灰度值之和,该处即为静脉纹线处。该方向的模板不但避免了阈值选择,能够排除对比度低、光照不均匀的影响,而且可以有效抑制指节纹等干扰纹线。实验结果表明,该方法可以有效地解决指节纹干扰、对比度低和光照不均等问题,提取的静脉纹线具有很好的连续性。

Due to the problem of poor effect of finger vein segmentation under the condition of uneven illumination,lowcontrast and knuckles lines interference,a kind of finger vein detection method based on local gray minimum is put forward. According to the direction of the finger vein lines,a template perpendicular to the direction of finger vein is selected which consists of sub-templates. Because gray value in vein is lower than that of its surrounding,when inspection template makes the point to point detection from top to bottom,the sum of gray value for middle sub-templates is less than that of the other two,which is the vein lines. It not only avoids threshold selection and eliminates the influence of uneven illumination and lowcontrast,but also restrains knuckles lines interference effectively. The experiment shows that this method can solve the problem of knuckles lines interference,lowcontrast and uneven illumination and extract finger vein with good continuity.