快速缝翘曲图像重定向算法  

Fast seam warping image retargeting algorithm

作　　者：赵亮 杜振龙 彭宏京 ZHAO Liang;DU Zhen-long;PENG Hong-jing(College of Computer Science and Technology,Nanjing Tech University,Nanjing 211816,China)

机构地区：[1]南京工业大学计算机科学与技术学院,江苏南京211816

出　　处：《计算机工程与设计》2021年第1期167-174,共8页Computer Engineering and Design

基　　金：国家自然科学基金项目(61672279);江苏省“六大人才高峰”基金项目(2012-WLW-023)。

摘　　要：针对现有适用于小型显示设备的图像重定向算法焦点区域检测不精确、算法处理速度慢的问题,提出一种快速缝翘曲(FSW)图像重定向算法。建立一个与人类感知更加相关的自适应能量图,该能量图合并梯度幅度、深度图、显著性、面部、边缘和直线检测;提出一种利用行(列)像素之间最优匹配关系快速寻找最优高能量缝的方案实现快速、精确的焦点区域检测;通过给焦点和非焦点区域设置缩放因子以及使用线性插值方法,有效实现快速缝翘曲图像重定向。实验结果表明,该算法可更好保留复杂图像的重要内容及结构信息,其处理速度相较于其它算法也有明显提升。Aiming at the problems of inaccurate focus area detection and low algorithm processing speed of existing image retargeting algorithms suitable for small display devices,a fast seam warping(FSW)image retargeting algorithm was proposed.An adaptive energy map which was more related to human perception was established.The energy map was combined with gradient amplitude,depth map,salience,face,edge and line detection.A scheme for quickly finding the optimal high-energy seam using the optimal matching relationship between rows(columns)of pixels was also proposed to achieve fast and accurate focus area detection.By setting the scaling factors for the focus and non-focus areas and using the linear interpolation method,the fast seam warping image retargeting was realized effectively.Experimental results show that the proposed algorithm can better retain the important content and structure information of complex images and its processing speed is significantly improved compared with other algorithms.

关 键 词：图像重定向 焦点区域检测 最优匹配关系 寻缝方案 缝翘曲 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]