A Novel Approach to Fast Medical Image Fusion Based on Lifting Wavelet Transform

Medical image fusion as a new and promising research area is very useful in clinical diagnosis and treatment planning. In this paper, a new algorithm based on Lifting Wavelet Transform (LWT) to fuse multi-modality medical images is presented. The lifting scheme is a new idea of constructing wavelets and has several unique advantages in comparison with conventional convolution-based Wavelet Transform (WT). It allows for an in-place implementation of WT and reduces computation time and memory requirement greatly. So in this fusion algorithm, LWT is employed to decompose and reconstruct images to realize fast image fusion. Meanwhile, a new local feature-based fusion rule is put forward to improve fusion quality and extract all significant features from multi-source images. Extensive experiments including the fusion of registered medical CT/MRI, CT/SPECT, MRI/PET images are performed and mutual information is taken as a means of objective assessing fusion images. The experimental results show that the new approach is very effective and has a good performance to fuse medical images.