A maximum likelihood approach to joint groupwise image registration and fusion by a Student-t mixture model

In this paper, we propose a Student- t mixture model (SMM) to approximate the joint intensity scatter plot (JISP) of the groupwise images. The problem of joint groupwise image registration and fusion is considered as a maximum likelihood (ML) formulation. The parameters of registration and fusion are estimated simultaneously by an expectation maximization (EM) algorithm. To evaluate the performance of the proposed method, experiments on several types of multimodal images are performed. Comprehensive experiments demonstrate that the proposed approach has better performance than other methods.