结合VAM和模糊逻辑的NSCT图像融合方法

Image fusion based on NSCT using VAM and fuzzy logic

提出一种基于非下采样Contourlet变换(NSCT)的图像融合方法。NSCT分解具有平移不变性,有利于更好地保持原始图像的边缘信息和轮廓结构。由于图像融合任务的不确定性及模糊逻辑在处理该类问题时的优越性,在高低频的融合策略中引入模糊逻辑进行基于隶属度的融合。同时考虑到人类视觉感兴趣区域的重要性,在低频系数的融合中引入视觉注意机制,利用原始图像本身的显著区域信息来指导融合过程,从而在融合过程中最大限度地保留源图像中的显著信息。实验结果表明,算法的融合图像具有良好的视觉效果及客观评指标。

A method of image fusion based on Non-Subsampled Contourlet Transform（NSCT） is proposed.NSCT which is fully shift-invariant can maintain the edges information and the contour of original image.To handle the uncertainty of the contribution of the source images to the fused image,fuzzy logic is adopted to the fusion strategy.Considering of the impor- tance of interesting regions,Visual Attention Mechanism（VAM） is adopted to guide the fusion procedure by the salient information of the source images.The experimental results show the fused images of proposed algorithm have favorable visual effect and objective measures.