基于区域填充的图像修复算法研究  

Image Restoration Algorithm Based on Region Filling

作　　者：邓悟[1] 吴笛[1] 滕奇志[1] 宁宁[1] 何小海[1] 

机构地区：[1]四川大学图像信息研究所,成都610041

出　　处：《计算机与数字工程》2014年第3期495-498,525,共5页Computer & Digital Engineering

基　　金：国家自然科学基金(编号:61071161);四川省科技厅纵向项目(编号:2011SZ0208)资助

摘　　要：图像修复算法是一类基于人工智能研究领域的一个很有发展潜力和实用价值的研究方向。论文通过待修复区域中提取的信息,按照等照度线检索(isophote-driven)进行正交迭代运算,并在填充顺序中加入了判断策略,从而实现了对损坏图像的修复。通过对算法的分析和对比,验证了该算法的改进是有效的。Image restoration algorithm is a recently developed research direction based on artificial intelligence which has great development potential and practical value. To restore the impaired image, information is collected from the area to be restored. Then, the information is taken into the orthogonal iteration process using isophote-driven retrieval method. A determining policy is added to the filling process. After deep analysis and careful comparison, this algorithm is proved to be effective.

关 键 词：图像修复 图像补全 等照度线 洋葱皮策略 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]