符合人眼视觉特性的医学图像融合算法

针对医学图像的特点，基于人眼视觉特性，结合降噪算法提出一种图像融合方法，以提高融合图像的对比度，减小噪声对图像的干扰。算法根据噪声点和图像特征所具备的不同特性，对源图像进行降噪，过滤小波分解后对应的噪声系数。小波分解后的系数矩阵，通过计算各系数的局部梯度，以源图像区部梯度为判断依据建立算法，选择源图中不同方向上梯度大的系数作为最终融合系数。实验结果表明，该算法提升融合图像的信息量的同时，有效保护了图像的细节，改善了视觉效果。

Medical Image Fusion Algorithm Fitting Human Visual Properties

Aiming at the properties of medical images, This paper presents an algorithm based on human visual properties, with the use of de-noising algorithm, the algorithm improve the contrast of the fusion image and decrease the influence of noises. Based on the differences between noises and image characters, the algorithm decrease noises of source images first, the wavelet transform coefficients of noises will be removed. To the coefficients of high frequency, fusion rules using different oriental gradient of the coefficients to select fusion coefficients, bigger ones will finally construct the fusion image. The coefficients of low frequency will be given different weighting; image with big entropy has big weighting. The experimental results show that new algorithm improve the entropy of the image, protect the characters of the image and look well.