摘要
根据人体手指静脉图像的结构及特点,提出了一种能很好地抑制噪声、剔除图像的伪边缘、准确定位手指静脉图像边缘信息的方法。首先采用高频强调滤波对梯度图像进行增强,然后利用B样条函数对增强后的图像进行多次拟合,最后对拟合的曲面进行分水岭分割。实验结果表明,通过该方法处理的梯度图像再进行分水岭变换,有效避免了过分割问题,同时准确定位了手指静脉图像边缘信息,提高了手指静脉图像分割精度。
With regard to the structure and features of human finger vein image segmentation, a new algorithm that can effectively restrain noise, eliminate image edges and detect the finger vein image edges exactly was presented. Firstly, morphological gradient image edge was enhanced by high frequency emphasizing filter effectively. Then, the enhancement image is fit many times by B-spline function. Finally, watershed segmentation is used for the smoothing surface. Experimental results demonstrate that watershed segmentation with the proposed algorithm resolves the phenomena of over-segmentation well. The scheme can detect the finger vein image edges exactly and improve segmentation precision.
出处
《压电与声光》
CSCD
北大核心
2008年第2期233-235,共3页
Piezoelectrics & Acoustooptics
基金
重庆市教育委员会科学技术研究基金资助项目(KJ070620)
