张量投票在图像修复中的应用

Application of Tensor Voting in Image Inpainting

图像修复技术是计算机图形学和机器视觉中的研究热点之一,在文物保护、多余目标物体剔除、影视特技制作、图像缩放、图像的有损压缩、视频通信的错误隐匿等方面均有重大应用价值。在分类总结相关工作的基础上,介绍了张量投票方法在图像修复中的应用,并展示了该方法的实验处理效果。

Image inpainting is one of research hotspots both in computer graphics and computer vision, which can be applied in historic preservation, unwanted objects elimination, film and television special effects production, image resizing, image compression and error concealment of video communication. This paper presents a summary of related works and introduces how to apply the tensor voting method in optical pre-processing. The algorithm’s flowchart and typical experimental result are demonstrated to show the completion characteristics of the algorithm.