摘要
本文讨论了一种新的血管造影图像边缘检测方法。这一方法是首先调整正规化滤波器的正规化参数用以平滑血管造影图像,去除低于这参数尺度的细节噪声,再利用小波变换提取血管的边缘。我们的结果表明,在相当大的正规化参数范围内,有效地去除不需要的细节噪声的同时,还能很好的保证提取的血管边缘的精度。与单独使用小波变换或递推正规化滤波器相比,该方法的优点是既弥补了小波变换去除相关噪声差的弱点,又改善了正规化滤波器局部分辨率差的缺点。
This paper discusses a new method by which the boundaries of the objects of the angiography can be located with recursive regularization filter and wavelet transform.The edges of the vessel can be accurately abstracted by the wavelet transform,after the noise and objects that fall below a scale of interest have been selectively removed by the filter under the regularization extent chosen.The result in this paper has explicitly show that it provides adequately accurate boundaries of the vessel that are of sufficient size in the width regularization extent.Comparing the method directly using wavelet transform and regularization filter respectively with the method which we proposed,the advantage of the method describing in this paper is complementary to the default that the wavelet transform badly reduces the correlative noise and add the multiresolution analysis which the regularization filter do not share.
出处
《医用生物力学》
CAS
CSCD
1995年第4期209-213,248,共6页
Journal of Medical Biomechanics
基金
国家自然科学基金