跳到主要导航 跳到搜索 跳到主要内容

A novel image fusion algorithm based on structural similarity

  • Xi'an Jiaotong University

科研成果: 书/报告/会议事项章节会议稿件同行评审

1 引用 (Scopus)

摘要

In this paper, we propose an improved image fusion algorithm, which uses structural similarity to match the local energy of images. The local energy method can reflect the complemen-tarity between different types of images. But for grayscale images and infrared images which have very different intensity, the fusion between them can produce dark spots in the result. To avoid this, we introduce a structural similarity model. We also improve the low and high frequency fusion rules to avoid the loss of low frequency information and keep the edge information in the image. We did experiments in Matlab and the results show that our method out-perform the traditional methods with respect to entropy, average gradient and spatial frequency.

源语言英语
主期刊名Proceedings - 2017 2nd International Conference on Multimedia and Image Processing, ICMIP 2017
出版商Institute of Electrical and Electronics Engineers Inc.
128-135
页数8
ISBN(电子版)9781509059546
DOI
出版状态已出版 - 15 12月 2017
活动2nd International Conference on Multimedia and Image Processing, ICMIP 2017 - Wuhan, Hubei, 中国
期限: 17 3月 201719 3月 2017

出版系列

姓名Proceedings - 2017 2nd International Conference on Multimedia and Image Processing, ICMIP 2017
2017-January

会议

会议2nd International Conference on Multimedia and Image Processing, ICMIP 2017
国家/地区中国
Wuhan, Hubei
时期17/03/1719/03/17

学术指纹

探究 'A novel image fusion algorithm based on structural similarity' 的科研主题。它们共同构成独一无二的指纹。

引用此