Unsupervised Deep Image Fusion With Structure Tensor Representations

Jung, Hyungjoo; Kim, Young-Jung; Jang, Hyunsung; Ha, Namkoo; Sohn, Kwanghoon

doi:10.1109/tip.2020.2966075

Cited by 116 publications

(30 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, we are explaining the conversion procedure of multispectral image to grayscale using the "Di Zenzo Structure tensor matrix". 37,38 Consider an image I in a rectangular grid format and expressed as = {1 … … . m} × {1 … … q}.…”

Section: : Calculate First-order Image Statistical Features For the mentioning

confidence: 99%

“…There are several methods described to separate luminance information from hue and saturation, for example, YIQ, HSV, LHS, CIELab, 35,36 and so on. The grayscale images store luminance information effectively within themselves, whereas the multispectral images are converted to grayscale by preserving the image “contrast.” Here, we are explaining the conversion procedure of multispectral image to grayscale using the “Di Zenzo Structure tensor matrix” 37,38 …”

Section: Proposed Algorithm Of Data Reduction and Privacymentioning

confidence: 99%

See 1 more Smart Citation

Conversion of Landsat 8 multispectral data through modified private content‐based image retrieval technique for secure transmission and privacy

Shakya

Ramola

Vidyarthi

2020

Engineering Reports

View full text Add to dashboard Cite

In this research work, we have developed a new image encryption-based algorithm for Landsat 8 satellite images. Image statistical parameters related to first-order image statistics, that is, mean (μ), SD (σ), and variance (σ 2) are used to obtain the feature values for both original images and their encrypted versions. Multispectral satellite images generally contain a wide range of bands, which is 9 to 11 specifically for the Landsat 8 satellite. Hence, the information content of these images is richer than typical images. Through this algorithm, we aim to reduce the image content by randomly selecting only 10% of the total image pixels. These pixels are used to create new images, the so-called encrypted images. The visual presentation of the encrypted image resembles a noisy picture. First-order image statistics are calculated for both original images and their encrypted counterparts. These parameters show a similarity in their numerical values for both original and encrypted images. Thus, noisy pictures can be used for transmission in place of the original image when an outcome is decided through statistical image features. The visual similarity in the original images and encrypted images is obtained from the histogram signature plot for both original and encrypted images. The proposed algorithm can hence be used to convert image data for secure transmission and privacy.

show abstract

Section: : Calculate First-order Image Statistical Features For the mentioning

confidence: 99%

Section: Proposed Algorithm Of Data Reduction and Privacymentioning

confidence: 99%

Conversion of Landsat 8 multispectral data through modified private content‐based image retrieval technique for secure transmission and privacy

Shakya

Ramola

Vidyarthi

2020

Engineering Reports

View full text Add to dashboard Cite

show abstract

“…For instance, long-time exposure will cause blurred edges, flashlight will introduce uneven shadows and overexposure. It is crucial to obtain clear images in some artificial intelligence tasks such as face recognition [7,8], target detection [9][10][11], image fusion [12] and autonomous driving [11,13].…”

Section: Introductionmentioning

confidence: 99%

AMBCR: Low‐light image enhancement via attention guided multi‐branch construction and Retinex theory

Zhou

Nie

et al. 2021

IET Image Processing

View full text Add to dashboard Cite

Due to different lighting environments and equipment limitations, low-light images have high noise, low contrast and unobvious colours. The main purpose of low-light image enhancement is to preserve the details and suppress noise as much as possible while improving the contrast of the image. Here, different networks are first combined to construct a multi-branch module for features extraction, and use the module and Retinex theory to extract the reflection map of the image. Then an attention mechanism is introduced into the multi-branch construction to balance the feature weight of each branch, and get the final result by the reconstruction module. The Retinex theory is used to calculate the L 1 loss and the gradient loss for the intermediate feature map of the entire model to train our framework. The entire process is completed in an end-to-end-way, which avoids the handcrafted reconstruction rules and reduces the workload. What's more, a large number of experiments demonstrate that the proposed framework performs better results than stateof-the-art algorithms in both quantitative and qualitative evaluations of image enhancement.

show abstract

“…Zhang et al [30] proposed a general image fusion framework named IFCNN based on a convolutional neural network. Jung et al [31] introduced a deep image fusion network (DIF-Net) as an unsupervised deep learning framework for image fusion. Ma et al [32] proposed using the generative adversarial network (GAN) for infrared and visible image fusion, achieving boosted results in detail retention.…”

Section: Introductionmentioning

confidence: 99%

Research on Infrared and Visible Image Fusion Based on Tetrolet Transform and Convolution Sparse Representation

et al. 2021

View full text Add to dashboard Cite

Image fusion is a visual enhancement technique that combines source images from different sensors to produce a more robust and informative fused image for subsequent processing or decision making. Infrared and visible light images share complementary properties that enable the production of robust and informative fused images. This paper proposed an infrared and visible image fusion method that improved the tetrolet framework to improve infrared and visible image fusion quality. First, the source image is enhanced by bicubic interpolation. The improved tetrolet transform then decomposes the enhanced source image; the high-frequency components are fused by convolutional sparse representation theory and combined with corresponding rules, and the low-frequency components are fused by defining ISER descriptors. Finally, we use the inverse transform to reconstruct the fused image. Qualitative and quantitative experimental results on five groups of typical infrared and visible image datasets demonstrate the proposed method's effectiveness. The proposed method exhibits better performances on subjective vision and objective indexes compared with the other state-of-the-art methods.INDEX TERMS Image fusion, improved tetrolet transform, convolutional sparse representation, ISER descriptor.

show abstract

Unsupervised Deep Image Fusion With Structure Tensor Representations

Cited by 116 publications

References 61 publications

Conversion of Landsat 8 multispectral data through modified private content‐based image retrieval technique for secure transmission and privacy

Conversion of Landsat 8 multispectral data through modified private content‐based image retrieval technique for secure transmission and privacy

AMBCR: Low‐light image enhancement via attention guided multi‐branch construction and Retinex theory

Research on Infrared and Visible Image Fusion Based on Tetrolet Transform and Convolution Sparse Representation

Contact Info

Product

Resources

About