Medical image fusion using regional statistics of shift-invariant shearlet domain

This paper presents a novel medical image fusion scheme for superior visualization by using shift-invariant shearlet transform (SIST) with regional statistics to preserve minimum color distortion and maintain the inner-dependency in different high pass frequencies. Firstly, SIST is used into each image. Then, we use Gamma probability distribution (GPD) to estimate adaptive threshold for selecting the best lowpass band automatically and modify highpass band by an adjust factor derived from highpass band. Finally, we obtain fusion results by the inverse transformation of SIST. Our experiments used PET and MRI medical images for performance evaluation. Experimental results show that the proposed method is feasible with the better quality than the state-of-the-art one.