作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 图形图像处理 • 上一篇    下一篇

基于视觉注意机制的异源图像融合

胡燕翔,万 莉   

  1. (天津师范大学计算机与信息工程学院,天津300387)
  • 收稿日期:2014-02-18 出版日期:2015-03-15 发布日期:2015-03-13
  • 作者简介:胡燕翔(1969 - ),男,副教授、博士,主研方向:图像处理,计算机视觉;万 莉,硕士研究生。
  • 基金资助:
    国家自然科学基金资助项目(61274021)。

Difference-source Image Fusion Based on Visual Attention Mechanism

HU Yanxiang,WAN Li   

  1. (College of Computer and Information Engineering,Tianjin Normal University,Tianjin 300387,China)
  • Received:2014-02-18 Online:2015-03-15 Published:2015-03-13

摘要: 针对融合规则不能真实反映观察者视觉感知特点的问题,提出一种异源图像多尺度融合算法。应用视觉 注意机制,计算异源图像间的视觉显著度匹配系数,并根据这一系数对小波变换近似系数进行自适应融合,在小波 细节系数融合中使用带有方向连续性检查的选大值方法,采用视觉显著度差对融合算法的视觉保持一致性进行评 价。测试结果表明,与传统算法相比,该算法在客观性能指标和主观视觉一致性方面都有所提高。

关键词: 异源图像融合, 视觉注意机制, 视觉一致性, 视觉显著度, 多尺度融合算法

Abstract: A Visual Attention Mechanism(VAM) based multi-scale image fusion algorithm is proposed in this paper, which tries to overcome the inconformity between fusion rules and observer′ s visual characteristics. Visual saliency matching coefficients based on visual attention mechanism are computed and employed in Wavelet Transform (WT) approximate coefficient fusion. In WT detail coefficient fusion,select-max rule with directional consistency checking is used. A fusion performance evaluating method based on the comparison of visual saliency is proposed to measure visual consistency between source images and fused results. Experimental results demonstrate the superiority of the proposed algorithm in terms of subjective visual similarity and objective quantization comparison.

Key words: different-source image fusion, Visual Attention Mechanism(VAM), visual consistency, visual saliency, multi-scale fusion algorithm

中图分类号: