Multi-Source Image Fusion Based on Low-Rank Decomposition and Convolutional Sparse Coding

Jiaxin Wang; Sheng Chen; Minghong Xie

doi:10.3788/LOP202158.2210009

Laser & Optoelectronics Progress, Volume. 58, Issue 22, 2210009(2021)

Multi-Source Image Fusion Based on Low-Rank Decomposition and Convolutional Sparse Coding

Jiaxin Wang¹, Sheng Chen², and Minghong Xie^1、*

¹Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming, Yunnan, 650500, China

²Gree Electric Appliances, INC. of Zhuhai, Zhuhai, Guangdong 519000, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(7)

Fig. 1. Framework of multi-source image fusion method based on low-rank decomposition and convolution sparse coding

Download full size

Fig. 2. Comparison of medical image results with different fusion methods. (a) CT image; (b) MRI image; (c) Ref. [26]; (d) Ref. [27]; (e) Ref. [28]; (f) Ref. [29]; (g) Ref. [30]; (h) proposed method

Download full size

Fig. 3. Comparison of infrared visible image results with different fusion methods. (a) Visible image; (b) infrared image; (c) Ref. [26]; (d) Ref. [27]; (e) Ref. [28]; (f) Ref. [29]; (g) Ref. [30]; (h) proposed method

Download full size

Fig. 4. Comparison of multi-focus image results with different fusion methods. (a) Far-focused image; (b) near-focus image; (c) Ref. [26]; (d) Ref. [27]; (e) Ref. [28]; (f) Ref. [29]; (g) Ref. [30]; (h) proposed method

Download full size

Table 1. Evaluation results of different fusion methods for medical images

View table

Table 1. Evaluation results of different fusion methods for medical images

Test image /(pixel×pixel)	Method	E	Q_NCIE	Q_M	Q_MI	Q_ab/f
CT/MRI(256×256)	Ref. [26]	4.9792	0.8081	0.4463	3.2700	0.5208
	Ref. [27]	4.9883	0.8081	0.2892	3.2163	0.5827
	Ref. [28]	5.2201	0.8079	0.3019	3.1763	0.5021
	Ref. [29]	4.9933	0.8080	0.2669	3.2535	0.4392
	Ref. [30]	5.3853	0.8074	0.3040	2.9723	0.5243
	Proposed method	5.3761	0.8122	0.8579	4.1977	0.6060

Table 2. Evaluation results of infrared visible images by different fusion methods images

View table

Table 2. Evaluation results of infrared visible images by different fusion methods images

Test image /(pixel×pixel)	Method	E	Q_NCIE	Q_M	Q_MI	Q_ab/f
Street(256×256)	Ref. [26]	6.1644	0.8050	1.1850	2.2085	0.5958
	Ref. [27]	6.6830	0.8064	0.3820	2.5973	0.6090
	Ref. [28]	5.8942	0.8047	0.4533	2.1007	0.4558
	Ref. [29]	6.1803	0.8048	0.7416	2.1433	0.5880
	Ref. [30]	6.4370	0.8045	0.8281	2.0249	0.6310
	Proposed method	6.7780	0.8164	0.9619	4.3619	0.6763

Table 3. Evaluation results of multi-focus images by different fusion methods

View table

Table 3. Evaluation results of multi-focus images by different fusion methods

Test image /(pixel×pixel)	Method	E	Q_NCIE	Q_M	Q_MI	Q_ab/f
clock (256×256)	Ref. [26]	7.3224	0.4253	1.8322	7.5078	0.7323
	Ref. [27]	7.3911	0.8283	0.7370	6.6713	0.6890
	Ref. [28]	7.3706	0.8292	0.9837	6.8377	0.6865
	Ref. [29]	7.3703	0.8329	1.3995	7.3892	0.7191
	Ref. [30]	7.3649	0.8317	1.8445	7.1928	0.7265
	Proposed method	7.3306	0.8357	1.0500	7.8026	0.6629

Tools

Get Citation

Copy Citation Text

Jiaxin Wang, Sheng Chen, Minghong Xie. Multi-Source Image Fusion Based on Low-Rank Decomposition and Convolutional Sparse Coding[J]. Laser & Optoelectronics Progress, 2021, 58(22): 2210009

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Image Processing

Received: Nov. 19, 2020

Accepted: Feb. 4, 2021

Published Online: Nov. 5, 2021

The Author Email: Minghong Xie (minghongxie@kust.edu.cn)

DOI:10.3788/LOP202158.2210009

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology