基于MRF和图割的多尺度图像分割和配准同步方法

Multi-scale joint segmentation and registration of image based on MRF and graph cut

图像分割和图像配准是图像处理领域中的两个关键技术.事实上,图像分割和图像配准之间是相互依赖、彼此互惠的.本文以图割理论为基础,结合马尔科夫随机场(Markov Random Field,MRF)模型,提出了一种基于MRF和图割的多尺度图像分割和配准同步方法,该方法通过将分割信息和配准信息耦合在一起,利用两者的相互促进作用,以得到更准确的分割和配准结果,并通过实验证明了该方法的有效性.

Image segmentation and registration are fundamental tasks in the field of image processing. Indeed, image registration significantly benefits from previous segmentation and vice versa. In this paper, a multi-scale joint segmentation and registration approach based on the theory of graph cut and the MRF model is presented. The method couples segmentation information and registration information together, making segmentation and registration both have a higher accuracy. Experimental results verify the effectiveness and the efficiency of the proposed method.