| 特征向量一维化的图像修复算法 |
| |
| 引用本文： | 胡晓彤,曹东瑞. 特征向量一维化的图像修复算法[J]. 计算机仿真, 2008, 25(3): 231-234 |
| |
| 作者姓名： | 胡晓彤 曹东瑞 |
| |
| 作者单位： | 天津科技大学,天津,300222;天津科技大学,天津,300222 |
| |
| 摘 要： | 图像修复是指对图像中的缺损部分进行自动修复.图像结构特征向量一维化的图像修复算法,将图像转换为一维图像结构特征向量后,构建图像结构特征空间,然后在该特征空间内寻找一组与每一个含有缺损部分的图像区域高度相似的特征向量,并利用这组特征向量在特征空间内定义一子空间,最后基于该特征子空间修复缺损像素.试验表明,经过图像的一维化后,增加了特征向量的数量,从而能够更加准确地修复缺损区域.
|
| 关 键 词： | 图像修复 自相似性 特征向量一维化 图像结构特征空间 |
| 文章编号： | 1006-9348(2008)03-0231-04 |
| 修稿时间： | 2007-03-14 |
Image Inpainting Based on One-dimensional Eigenvector for Image Structure |
| |
| HU Xiao-tong,CAO Dong-rui. Image Inpainting Based on One-dimensional Eigenvector for Image Structure[J]. Computer Simulation, 2008, 25(3): 231-234 |
| |
| Authors: | HU Xiao-tong CAO Dong-rui |
| |
| Affiliation: | HU Xiao-tong,CAO Dong-rui(Tianjin University of Science , Technology,Tianjin 300222,China) |
| |
| Abstract: | The image inpainting is the technology to restore the damaged part in image automatically. This paper proposes a new method based on one-dimensional eigenvector for image structure. The proposed method scans images into one-dimensional eigenvector firstly, and then, searches different set of image blocks that have high similarity with each damaged image block, to make up eigensubspace for image structure. At the end, the damaged image blocks can be restored according to their eigensubspace. Experiment resul... |
| |
| Keywords: | Image inpainting Self-similarity One-dimensional eigenvector Eigenspace for image structure |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
