Super-resolution image reconstruction method for any blurred kernel
A low-resolution image and image reconstruction technology, applied in the field of super-resolution reconstruction of low-resolution images, can solve problems such as unformed systematic results
Pending Publication Date: 2020-12-11
XIAN UNIV OF SCI & TECH
View PDF0 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, there are various super-resolution image reconstruction algorithms based on deep learning, and the research on reconstruction of low-resolution images with arbitrary blur kernels is still in its infancy, and no systematic results have been formed.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0039] Such as figure 1 As shown, the visual reconstruction of the present invention includes the following steps:
[0040] Before starting all improvement work, design a degradation model suitable for images with arbitrary blur kernels;
[0041] Furthermore, this degradation model not only takes into account the shortcomings of the existing general degradation models, but also considers the shortcomings of the bicubic degradation model being too simple. Based on the existing two most common degradation models, the respective After analyzing the advantages and disadvantages, a new degradation model is obtained;
[0042] Further, the mathematical expression of the degradation model is:
[0043]
[0044] where y is a low-resolution (LR) image, x is a high-resolution (HR) image, k is a blur kernel, G is white Gaussian noise (AWGN) at a certain noise level, ⊕ is a convolution operation, ↓ s " is the downsampling operation of scale factor s.
[0045] Further, in order to so...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a depth plug-and-play super-resolution image reconstruction method for any blurred kernel based on deep learning, and the method can achieve the real-time super-resolution reconstruction of a low-resolution image based on any blurred kernel. The method comprises a shallow feature extraction module; a deep feature extraction module; an image up-sampling module; an image reconstruction module; and an evaluation module. The shallow feature extraction module performs shallow extraction of high and low frequency feature information in the input low-resolution image; the deepfeature extraction module performs deep extraction of high and low frequency feature information in the input high-resolution image; the image up-sampling module amplifies the deep feature information; and the image reconstruction module reconstructs the amplified information after mapping, and the evaluation module performs performance evaluation on the reconstructed image and the original high-resolution image. According to the invention, super-resolution image reconstruction can be carried out on an input low-resolution image with any blurring kernel under different magnification ratios.
Description
technical field [0001] The invention relates to the technical field of low-resolution image super-resolution reconstruction in image processing, in particular to a super-resolution image reconstruction method for arbitrary blur kernels. Background technique [0002] More than 80% of the information that humans receive from the outside is visual information. As a basic part of the interaction between humans and the environment, vision is generally stored in the form of images or videos. Relying on images that carry visual information, technologies such as object detection have emerged as the times require. However, due to hardware cost and environmental constraints, images are generally stored and displayed at lower resolutions. In order to enable images to be displayed in a high-resolution form, the problem is generally solved from two levels of hardware and software. Practice has proved that compared with solving problems from the hardware level, solving problems from the...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06T3/40G06T5/00
CPCG06T3/4053G06T2207/20081G06T2207/20084G06T5/73
Inventor 温帆
Owner XIAN UNIV OF SCI & TECH