摘要
医学图像融合在医学图像分析和诊断上具有极为重要的应用价值。本文针对小波变换的特点,对传统的小波融合算法进行改进,提出了一种新的确定融合图像高频分量和低频分量的方法。选择高频系数时,通过计算各子图像的区域边界强度来实现自适应的高频信息融合;低频系数采用基于边缘细节的融合算法,可以使融合的图像细节更丰富、更清晰。文中采用传统小波算法和改进的小波算法对人体的两组数据进行融合处理,并给出了量化的评判指标,实验评价参数表明,本文提出的图像融合技术能有效地保留原始图像的细节特征,强化两幅图像中的边缘和纹理特征,优于传统小波融合方法。
Medical image fusion is of very important value for application in medical image analysis and diagnosis. In this paper, the conventional method of wavelet fusion is improved, so a new algorithm of medical image fusion is presented and the high frequency and low frequency coefficients are studied respectively. When high frequency coefficients are chosen, the regional edge intensities of each sub-image are calculated to realize adaptive fusion. The choice of low frequency coefficient is based on the edges of images, so that the fused image preserves all useful information and appears more distinctly. We apply the conventional and the improved fusion algorithms based on wavelet transform to fuse two images of human body and also evaluate the fusion results through a quality evaluation method. Experimental results show that this algorithm can effectively retain the details of information on original images and enhance their edge and texture features. This new algorithm is better than the conventional fusion algorithm based on wavelet transform.
出处
《生物医学工程学杂志》
EI
CAS
CSCD
北大核心
2009年第4期711-715,共5页
Journal of Biomedical Engineering
基金
国家自然科学基金资助项目(30671997)
关键词
图像融合
小波变换
融合算法
效果评价
Image fusion
Wavelet transform
Fusion algorithm
Quality evaluation