关键词： 图像配准, 空间自适应, 正则化参数, 不动点方法

Abstract: A discrete spatially adaptive total variation image registration model is established by introducing spatial adaptive regularization parameters, and an effective fixed point method is given by combining the first-order Taylor expansion. Experimental results show that the image obtained from the spatial adaptive total variation registration model has better quality and requires less iterative time than the total variation registration model.

Key words: image registration, spatially adaptive, regularization parameter, fixed point method