Image fusion based on wedgelet and wavelet

Wavelet transform has been widely used in image fusion; it can well represent smooth regions and regular textures. However, wavelet is unable to provide a good representation of 2D edges. Wedgelet transform has good performance in approximation to edges. In this paper, we propose a new fusion method; it makes good approximation to edges in original images by using wedgelet and retains texture features in residual images with wavelet. Through combining wedgelet with wavelet, we develop a detector of edges and texture features in residual images. So the respective superiorities of wedgelet and wavelet can be combined by adding the fusion results using wavelet on the fusion results using wedgelet respectively. The experiments show that this novel method provides improved qualitative results compared to previous wavelet and dual tree complex wavelet (DT-CWT) fusion methods.