采用树状小波变换的医学图像融合方法及实现

摘要

提出基于树状小波变换的医学图像融合方法,所提出的算法能够在一定的能量准则下对图像进行自适应树状小波分解,能够根据图像的子图能量决定图像的变换.通过实验仿真及客观的图像融合评估准则,实验结果表明了树状小波分解能够得到比塔形小波分解更好的融合效果,信噪比提高了很多,均方差明显减少,且该算法不但保留了小波变换算法的多分辨率特性,而且充分保留了图像中的细节信息,减少了融合的复杂度.%An algorithm for medical image fusion based on tree-structure wavelet transform is proposed. Following some special energy criterion, it can implement adaptive tree-structure wavelet decomposition on image and can determine image transform according to the sub-image energy of the image. The algorithm is simulated through experiment and assessed based on objective image fusion criteria. Experimental results show that the tree-structure wavelet decomposition can achieve better fusion outcome than that of the pyramid wavelet decomposition, the signal-to-noise ratio has improved a lot, and the mean-square deviation reduces significantly. The proposed algorithm not only retains the multi-resolution characteristics of the wavelet, but also fully retains the details in the image, and reduces the complexity of fusion.