基于罗伯茨梯度的彩色图像缩放新法  

Novel color image resizing based on Roberts gradient

作　　者：侯发忠[1,2] 邹北骥[1] 李跃强[2] 

机构地区：[1]怀化医学高等专科学校,湖南怀化418000 [2]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《计算机工程与设计》2009年第14期3367-3370,共4页Computer Engineering and Design

摘　　要：为了缩放图像而不改变图像中重要特征物的大小和比率,提出一种叫做缝络删增的缩放方法。一个缝络就是从顶到底或从左到右8-连接且仅连接一个像素的路径。最佳路径由图像的罗伯茨梯度函数来定义。通过在行列一个或二个方向重复删除或插入缝络就能改变图像的尺寸和比率。算法定义的缝络绕过了特征物边界,从而保护了图像的内容。理论分析和实验结果表明,该算法较其它算法能更有效地保护图像特征物。To resize image without changing the important characteristic＇ s size and rate, a simple resizing operator called seam carving is presented. A seam is an optimal 8-connected path ofpixels on a single image from top to bottom, or left to right, where optimality is defined by Roberts gradient function. By repeatedly carving out or inserting seams in one or both directions can change the aspect ratio and size of an image. The selection and order of seams Bypass the border of the characteristics, so protect the content of the image, theoretical analysis and experimental results show that this algorithm than other similar algorithms to provide more effective protection of image features.

关 键 词：图像 缩放 能量 函数 图像缝络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]