中文摘要：

针对现有多模态医学图像融合方法间的优点不易综合的问题，提出一种基于差分进化算法的多模态医学图像融合。分别利用ISH和小波变换方法产生初始图像，利用差分进化算法来优化由图像定量评价指标构成的目标函数，从而获取视觉效果较好，清晰度较好和信息量丰富的融合图像。实验结果和评价参数表明，提算法在平均梯度、熵等表示细节信息和融合质量的等定量指标上比基于小波变换和ISH变换图像融合算法提高了30％和2．5％以上，有效地结合基于小波变换和ISH变换两种不同融合方法的优点。

英文摘要：

It was hard to integrate advantages of the existing multi-modality medicine image fusion methods. In order to overcome this limitation, a multi-modality medicine image fusion method was proposed based on differential evolution （DE） algorithm. The initial images were obtained by employing the ISH and wavelet transform. The objective function was composed of image quantitative evaluation indices, and DE was employed to obtain a fusion image of the good visual effect and definition as well as abundant information. The experiment results and the evaluation index indicate that the new calculation method raises the ratio of quantitative indexes up to 30%（wavelet transform） and 2.5% （the ISH）, respec- tively, compared with the image fusion methods. Therefore, this method combines the merits of different fusion methods based on the ISH and wavelet transform.