摘要
目前的非规则缺失图像修复方法,连接图像的非规则缺失对角元素,导致修复图像所需时间长,修复后的图像的信噪比较低,修复误差较大,只能修复规则缺失图像,为此提出基于变分自编码器的非规则缺失图像修复仿真。利用变分自编码器,提取非规则缺失图像局部信息,预测图像非规则缺失部分,完成非规则缺失图像缺失部分填充;采用循环矩阵和特普利兹矩阵,重建非规则缺失图像;设计非规则缺失图像修复仿真流程,完成非规则缺失图像修复。实验结果表明,所设计的非规则缺失图像修复方法,修复图像中存在的不规则缺失,产生的信噪比较高、修复误差较小,修复所需时间较短,具有较高的不规则图像缺失修复性能,较优的图像修复效果。
In the process of irregular image restoration, the lack of diagonal elements leads to long repair time,low signal-to-noise ratio and large repair error. Therefore, we reported the simulation of irregular missing image restoration based on variational self-encoder. The variational self-encoder was used to extract the local information of the irregular missing image, predict the irregular missing part of the image, and complete the filling of the missing part of the irregular missing image;The irregular missing image was reconstructed by using cyclic matrix and Teplitz matrix;The simulation process of irregular missing image restoration was deesignd and the irregular missing image restoration was completed. The experimental results show that this method has high signal-to-noise ratio, small repair error, short repair time and excellent repair effect.
作者
胡秋生
胡璋
HU Qiu-sheng;HU Zhang(Nanchang University College of Science and Technology,Gongqingcheng Jiangxi 332020,China)
出处
《计算机仿真》
北大核心
2021年第12期155-159,共5页
Computer Simulation
基金
江西省教育厅科学技术研究项目(191567)。
关键词
变分自编码器
非规则
缺失图像
修复仿真
Variational auto encoder
Irregular
Missing image
Repair simulation
