Color Image Registration Model Based on Abstract Matching Flow

A color image registration model within the framework of abstract matching flow is proposed to deal with the problem of serious color difference between registered images.The model is composed of a data term and a regularization term.The data term employs the cross correlation as the similarity measurement between two images,in order to cope with the large color difference problem.The regularization term into which an anisotropic flow-driven diffusion is introduced,aims at preserving image features during the image evolution.The diffusion structure tensor in the diffusion coefficient function is a vector-valued structure tensor,which integrates the intensity and structure information of each channel,as well as the correlation among channels.Experimental results validate the proposed model,especially for images of large color difference.