A novel image fusion approach based on compressive sensing

Yin, Hongpeng; Liu, Zhaodong; Fang, Bin; Li, Yanxia

doi:10.1016/j.optcom.2015.05.020

Cited by 17 publications

(8 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hybrid model-based methods aim to synthetically use the advantages of each independent method and avoid the effects of the disadvantages as possible. Different MSD hybrid methods [42,43], hybrid MSD and SR methods [18,44,45], hybrid MSD and neural networks methods [46,47], also hybrid MSD, neural network, and SR methods [48,49] all make great contributions to the research. Other fusion methods include, for example, intensity-hue-saturation (IHS) transform methods [50,51], fuzzy theory method [52], independent component analysis (ICA) method [53], and so on.…”

Section: Conventional Fusion Methodsmentioning

confidence: 99%

Infrared and Visible Image Fusion with a Generative Adversarial Network and a Residual Network

Wang

et al. 2020

Applied Sciences

View full text Add to dashboard Cite

Infrared and visible image fusion can obtain combined images with salient hidden objectives and abundant visible details simultaneously. In this paper, we propose a novel method for infrared and visible image fusion with a deep learning framework based on a generative adversarial network (GAN) and a residual network (ResNet). The fusion is accomplished with an adversarial game and directed by the unique loss functions. The generator with residual blocks and skip connections can extract deep features of source image pairs and generate an elementary fused image with infrared thermal radiation information and visible texture information, and more details in visible images are added to the final images through the discriminator. It is unnecessary to design the activity level measurements and fusion rules manually, which are now implemented automatically. Also, there are no complicated multi-scale transforms in this method, so the computational cost and complexity can be reduced. Experiment results demonstrate that the proposed method eventually gets desirable images, achieving better performance in objective assessment and visual quality compared with nine representative infrared and visible image fusion methods.

show abstract

Section: Conventional Fusion Methodsmentioning

confidence: 99%

Infrared and Visible Image Fusion with a Generative Adversarial Network and a Residual Network

Wang

et al. 2020

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…Its core idea is to reconstruct from indirect measurements with sampling signals at a smaller rate compared to conventional Shannon's sampling rate (Yin and Liu et al, 2015). In this work, the component of sensing measurement is discussed and conducted to measure the activity level of subbands coefficients.…”

Section: Block-based Random Image Samplingmentioning

confidence: 99%

“…In practical applications, the equivalent condition of RIP is that is incoherent with (Yin and Liu et al, 2015;Candès and Wakin, 2008). There are many matrices meeting the condition including Gaussian random matrix, Bernoulli matrix and Hadamard matrix, in which Gaussian random matrix is used in this work.…”

Section: Block-based Random Image Samplingmentioning

confidence: 99%

“…There are many matrices meeting the condition including Gaussian random matrix, Bernoulli matrix and Hadamard matrix, in which Gaussian random matrix is used in this work. There are two reasons to utilize the Gaussian random matrix: firstly, the matrix is incoherent with any sparse basis and requires less number of measurements to acquire the salient information; secondly, it can obtain better results in inverse solution and restrict the mean error to project the sparse coefficients (Yin and Liu et al, 2015).…”

Section: Block-based Random Image Samplingmentioning

confidence: 99%

See 1 more Smart Citation

Fusion of Infrared and Visible Images Using Nonsubsampled Shearlet Transform and Block-based Random Image Sampling

2016

Technical Journal of the Faculty of Engineering

View full text Add to dashboard Cite

This paper presents a novel image fusion method based on nonsubsampled shearlet transform (NSST) and block-based random image sampling for infrared and visible images. In the method, the NSST is firstly performed on each of the source images to obtain their low frequency and high frequency subbands coefficients. Then the low frequency coefficients are fused by combing the block-based random image sampling and the sliding widow technique, while the high frequency coefficients are fused using classical absolute value maximum choosing rule. The fused image is reconstructed though performing the inverse NSST. The experiments are carried out on five pairs of infrared and visible images using five traditional fusion methods to verify the effectiveness and efficiency of the proposed method and by comparing the fused results visually and objectively it is demonstrated that the proposed fusion method is competitive or even superior to the other wellknown methods.

show abstract

“…The suggested method gives speed advantage in comparison to other transform based techniques. Yin et al [18] suggested a fusion procedure utilizing compressive sensing. The authors used non-subsampled contourlet transform to divide the source images.…”

Section: Introductionmentioning

confidence: 99%

A New Hybrid Multifocus Image Fusion Model Using Single Optimum Gabor Filter

Agrawal¹,

Panda²,

Kumari³

et al. 2019

RIA

View full text Add to dashboard Cite

The limited depth of field in optical lenses and camera leads to output images having nonuniform focus. Fusing the focussed regions from many images of the identical scene results in an output image with uniform focus. Generally, the methods suggested for image fusion (IF) suffers from computational complexity. In this context, we suggest a new hybrid multifocus IF model using a single optimum Gabor filter. Another important contribution of this paper is that the Gabor filter is capable of distinguishing the clear and the blurry pixels. The key concept is to decompose inputs into blocks. Each of the blocks/patches is convolved with the single optimum Gabor filter for extracting the Gabor energy feature vector. A new patch is created for fusion based on the Gabor energy feature value of each pixel in the patch. The parameters of the single Gabor filter are optimized using a relatively new optimization technique termed squirrel search algorithm (SSA). The application of optimal Gabor filter to the multifocus image fusion problem is new. The suggested technique is tested with standard images having focus on distinct objects. The outcomes reveal that the suggested technique provides improved performance, it outperforms state-of-the-art classic fusion approaches in both objective and subjective assessments.

show abstract

A novel image fusion approach based on compressive sensing

Cited by 17 publications

References 38 publications

Infrared and Visible Image Fusion with a Generative Adversarial Network and a Residual Network

Infrared and Visible Image Fusion with a Generative Adversarial Network and a Residual Network

Fusion of Infrared and Visible Images Using Nonsubsampled Shearlet Transform and Block-based Random Image Sampling

A New Hybrid Multifocus Image Fusion Model Using Single Optimum Gabor Filter

Contact Info

Product

Resources

About