北京航空航天大学学报 ›› 2012, Vol. ›› Issue (9): 1235-1239.

• 论文 • 上一篇    下一篇

细缝裁减和非均匀映射结合的图像自适应方法

曹连超, 司农, 高美琴, 毋立芳   

  1. 北京工业大学 电控学院, 北京 100124
  • 收稿日期:2011-10-14 出版日期:2012-09-30 发布日期:2012-09-27
  • 基金资助:
    国家自然科学基金资助项目(61040052); 北京市优秀人才资助项目(2009D005015000010)

Content aware image distance measure for joint use of seam carving and non-homogeneous warping

Cao Lianchao, Si Nong, Gao Meiqin, Wu Lifang   

  1. College of Electronic Information and Control Engineering, Beijing University of Technology, Beijing 100124, China
  • Received:2011-10-14 Online:2012-09-30 Published:2012-09-27

摘要: 采用将细缝裁减和非均匀映射相结合的图像尺寸自适应框架,提出了一种基于内容的图像重要信息变形的度量方法.首先提取原始图像的重要性像素点,利用细缝裁减去掉一条像素细缝后,相应的重要性像素点会被更新.对于图像中保留的重要性像素点,计算它们的子图像平均偏差(ADSI,Average Difference of Sub Images);对于被移除的重要性像素点,计算它们的平均丢失能量(ALE,Average Lost Energy).通过ADSI和ALE可计算出重要信息变形的度量函数(IIDF,Important Information Deformation Function)的值,通过分析IIDF的趋势得到细缝裁减的终止条件,然后改用非均匀映射方法(non-homogeneous warping)将图像自适应到目标尺寸.实验结果证明,新方法处理的结果图像重要区域变形较小,并且计算效率比较高.

Abstract: A content aware image distance measure for joint use of seam carving and warping was proposed. Firstly, important pixels in the original image were extracted. After a seam was removed, the important pixels were updated. For the preserved important pixels, the average difference of sub images (ADSI) was computed. For the removed pixels, the average lost energy (ALE) was computed correspondingly. Finally, the important information deformation function (IIDF) was defined by combining ADSI with ALE. If IIDF satisfies the designed terminating condition, seam carving was terminated and non-homogeneous warping was used to get the image of a target size. Experimental results show that the proposed approach can remove unimportant regions while incurring less deformation to important regions and also keeps the aspect ratio of important objects closer to that in the original image.

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发