上海大学学报(自然科学版) ›› 2012, Vol. 18 ›› Issue (4): 349-354.doi: 10.3969/j.issn.1007-2861.2012.04.004

• 通信与信息工程 • 上一篇    下一篇

基于图像稀疏表示视觉显著度计算的自适应尺度调整方法

韩彪,杨卫英,郑玉婷   

  1. (上海大学 影视艺术技术学院,上海 200072)
  • 出版日期:2012-08-30 发布日期:2012-08-30
  • 通讯作者: 杨卫英(1957~),女,副教授,研究方向为数字媒体技术. E-mail:yangweiying@staff.shu.edu.cn
  • 基金资助:

    上海大学创新基金资助项目(SHUCX112306)

Image Retargeting Based on Visual Saliency Using Sparse Coding

HAN Biao,YANG Wei-ying,ZHENG Yu-ting   

  1. (School of Film and TV Arts & Technology, Shanghai University, Shanghai 200072, China)
  • Online:2012-08-30 Published:2012-08-30

摘要: 图像适应是指将图像按照不同的纵横比进行拉伸和压缩,以适应不同显示比例的显示终端(如手机、掌上电脑(personal digital assistant,PDA)、投影仪)的一种技术.研究基于视觉显著度的图像适应,具体实现分为视觉显著度的计算和图像的加权拉伸与压缩两个步骤.采用加权的稀疏表达残差来表示视觉显著度.在图像的加权拉伸与压缩中,提出一种对横向与纵向使用不同加权变化的快速算法,在保证图像处理效果的前提下,达到快速图像适应处理的效果.仿真实验结果表明,该方法对于实验图像具有鲁棒性和有效性.

关键词: 加权拉伸, 视觉显著度, 图像适应

Abstract: Image retargeting is a technique of adapting images to different aspect ratio for various display terminals such as cellphone, personal digital assistant (PDA) and projector. This paper proposes a method to deal with the problem using visual saliency. The method includes two steps: computing a saliency map, and modifying the image to suit different aspect ratio. A novel visual saliency model is used. The model indicates the weighted sparse coding residual as visual saliency. In the step of weighted resizing, a fast algorithm is proposed. It is a fast image adaptive process with no obvious distortion. The experimental results indicate the robust and effectiveness of the proposed method with different images.

Key words: image retargeting, visual saliency, weighted resize

中图分类号: