Multi-exposure image fusion based on tensor decomposition
- PDF / 2,236,620 Bytes
- 19 Pages / 439.37 x 666.142 pts Page_size
- 86 Downloads / 210 Views
Multi-exposure image fusion based on tensor decomposition Shengcong Wu 1 & Ting Luo 2 & Yang Song 2 & Haiyong Xu 2 Received: 12 July 2019 / Revised: 10 May 2020 / Accepted: 27 May 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
In this paper, a multi-exposure image fusion (MEF) method is proposed based on tensor decomposition and saliency model. The main innovation of the proposed method is to explore a tensor domain for MEF and define the fusion rules based on tensor feature of higher order singular value decomposition (HOSVD) and saliency. Specifically, RGB images are converted to YCbCr images to maintain the stability of color information. For luminance channels, luminance patches of luminance images are constructed 3-order sub-tensors, and HOSVD is used to extract features of sub-tensors. Then, the sum of absolute coefficients (SAC) of weight coefficients are defined. Meanwhile, considering the impact of saliency on visual perception, visual saliency maps (VSMs) is used to evaluate luminance patches quality and guide the fusion rules to define the rule of fusion. For chrominance channels, VSMs of the chrominance channels is used to define fused rule. The experimental results show that the fused image with more texture details and saturated color is successfully generated by proposed method. Keywords Multi-exposure image fusion . Tensor decomposition . Higher order singular value decomposition
* Haiyong Xu [email protected] Shengcong Wu [email protected] Ting Luo [email protected] Yang Song [email protected] Extended author information available on the last page of the article
Multimedia Tools and Applications 19
1 Introduction Due to the dynamic range limitation of digital cameras, images captured by cameras is often not as large as the dynamic range of natural scenes, and the details are seriously lost. Therefore, high dynamic range (HDR) imaging technique has become a hot topic of research, and has been widely used in many fields, such as data hiding and medical imaging [2, 6]. Generally, HDR imaging techniques estimate the camera response function (CRF) from multiple low dynamic range (LDR) images, and then uses CRF inverse operations to reconstruct HDR images [21, 23]. After acquiring the HDR image, it is necessary to use the tone mapping technique [7] to compress the dynamic range of the HDR image for display. However, the computational complexity of the tone mapping technique is high, and the quality of the HDR image is dependent on the computational accuracy of CRF. On the contrary, MEF methods use a series of LDR images to fuse high quality LDR image without CRF recovery and tone mapping, and has more information than any source input image. Therefore, the MEF method is an efficient alternative to complex HDR imaging technique [3, 29]. Recently, many MEF methods have been proposed and achieved performance well. Mertens et al. [19] proposed a MEF method by using three quality measures of contrast, saturation and well-exposedness to evaluate quality of pix
Data Loading...
