基于小波变换的多模态医学图像的融合及性能评价

为了研究小波变换分解的尺度和融合策略对图像融合效果的影响。我们选择已配准后的多聚焦医学图像以及MRI/CT灰度图像,在提取图像的低频和高频小波系数时,分别进行单尺度和多尺度分解,融合时采取了基于独立像素点和基于邻域窗口的多种融合策略,深入对比分析各种融合规则对医学图像融合性能的影响。实验结果和性能评价表明:使用局部滤波的操作可以明显改善图像融合的效果,使图像的细节信息更加丰富,而多尺度融合能明显提高融合图像的亮度。

Multi-morphological Fusion of Medical Image and Performance Evaluation based on Wavelet Transform

We choosed multi-focus images and standard MRI / CT gray images for source images and applied the following strategies to decomposition and fusion. On the one hand,decomposition of low frequency and high frequency wavelet coefficients was adopted with single scale and multi-scale way. On the other hand,image fusion region was selected by independent pixel and neighborhood operation. Through above different operation of decomposition and fusion,the fusion image of different fusion rule was obtained. Based on the fusion image of different fusion rule,the influences of various fusion rules on the fusion results were compared and analyzed,which was the purpose of this paper. By means of analyzing the experimental results and evaluation results of multi-focus fusion images and clinical MRI / CT fusion images. It show that neighborhood filtering operation can greatly improve the result of image fusion and make the detail of the image information more rich; multi-scale decompose can increase the brightness of the fusion image with respect to single scale decompose.