基于MRF模型的图像修复算法

Image Inpainting Algorithm Based on MRF Model

基于马尔可夫随机场(MRF)模型,提出一种直接采样的图像修复方法。通过优先权机制来决定边界像素的修复次序,充分顾及到图像的边缘结构信息,同时能够恢复纹理细节,避免了模糊效应。实验证明该算法具有很好的修复效果。

Based on MRF model, an inpainting algorithm using direct sampling was proposed. The inpainting order of the boundary pixels was determined by priority. Fully considering the edge features, the algorithm could also restore the texture details, and thus, blurring artifact was avoided. The experiments indicated the algorithm could achieve fairly good results.