Image interpolation method and system based on prediction gradient

A technology of interpolation method and prediction method, which is applied in image data processing, graphics and image conversion, instruments, etc., and can solve problems such as inconsistency, poor picture quality, and long time consumption

Active Publication Date: 2016-06-15
SUZHOU UNIV
View PDF2 Cites 21 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method is highly related to the trained dictionary database, and the results of image interpolation with different content will be different
As a result, when the original image image is not self-similar or no similar image patch is found in the dictionary database, the quality of the interpolated image will be poor and some artifacts will be produced
[0010] If the database is large, the offline training process and image interpolation online search process will take a long time, and the required computer memory is also a problem that needs to be solved

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image interpolation method and system based on prediction gradient
  • Image interpolation method and system based on prediction gradient
  • Image interpolation method and system based on prediction gradient

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0095] The image interpolation method based on the prediction gradient in this embodiment includes:

[0096] Determine whether the pixel to be interpolated is an edge pixel,

[0097] If the pixel to be interpolated is a non-edge pixel, a linear interpolation method is used for interpolation to obtain a high-resolution image;

[0098] If the pixel to be interpolated is an edge pixel, a non-linear interpolation method is used for interpolation to obtain a high-resolution image.

[0099] In this embodiment, given a continuous tone gradient image I(i, j), the gradient image can be regarded as a two-dimensional function about (i, j). In order to judge whether a certain pixel is an edge pixel, the present invention needs to calculate the gradient value of this pixel, and the calculation formula is as shown in formula (1):

[0100] Formula 1)

[0101] here, Refers to the gradient value on the pixel (i, j), is the directional derivative of the LR image, θ represents the direc...

Embodiment 2

[0105] In this embodiment, the image interpolation method based on the prediction gradient, on the basis of Embodiment 1, the nonlinear interpolation method specifically includes the following contents.

[0106] The greater the edge contrast, the easier it is to be observed by the human eye. The reason why the human visual system can observe the edge of a natural gradient image is because of the large gradient difference between adjacent regions, such as Figure 4 As shown in (a), the difference between such gradients can be referred to as contrast or edge intensity in the present invention. Using an edge detection algorithm, for Figure 4 (a) Given an image, one will get Figure 4 In the edge result graph shown in (b), the edge line formed at this time is the real edge of the gradient image, and it is also the edge observed by the human eye. But it is worth noting that the greater the edge strength, the easier it is to be regarded as an edge by the human eye. Unfortunately...

Embodiment 3

[0126] The image interpolation system based on the prediction gradient in this embodiment includes:

[0127] The edge pixel judging module is used to judge whether the pixel to be interpolated is an edge pixel,

[0128] If the pixel to be interpolated is a non-edge pixel, the linear interpolation module uses a linear interpolation method to perform interpolation to obtain a high-resolution image;

[0129] If the pixel to be interpolated is an edge pixel, the nonlinear interpolation module performs interpolation using a nonlinear interpolation method to obtain a high-resolution image.

[0130] In this embodiment, the performance of the same color image in different color spaces is different. Among them, the RGB color space is the most important and the most widely used color space, and the other color spaces are linear or nonlinear changes of the RGB color space. In the color space represented by RGB, a color picture can be expressed as the following model:

[0131]

[01...

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

No PUM Login to view more

Abstract

The invention discloses an image interpolation method and system based on a prediction gradient, and the method can effectively determine whether a pixel is an edge pixel and is designed to effectively improve the interpolation quality. The method comprises: whether a pixel to be interpolated is an edge pixel is determined; if the pixel to be interpolated is a non-edge pixel, a linear interpolation method is used for interpolation to obtain a high-resolution image; or if the pixel to be interpolated is an edge pixel, a non-linear interpolation method is used for interpolation to obtain a high-resolution image. The method can effectively determine image gradient values, keeps the edges and texture information, and reduces sawtooth phenomena and fuzzy phenomena in an interpolation algorithm, so that not much calculation time is taken to process a color image converted to a Lab color space, which is of great importance to real-time image application.

Description

technical field [0001] The invention relates to an image interpolation method and system based on prediction gradient. Background technique [0002] The spatial resolution of an image is an important feature for the human visual system to observe digital images. The higher the resolution, the more image detail information the human eye can obtain. High-resolution digital images are required in many fields, such as face recognition, military radar image processing, biomedical slice analysis, and digital high-definition television applications. Generally speaking, there are two ways to improve the image resolution, one is to upgrade the image acquisition equipment, and the other is to use software to improve the spatial resolution of the collected digital images. Although in the past few decades, with the development of digital camera sensor equipment, camcorders or scanners have made great progress, it is still essential to use software to improve the resolution of digital i...

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
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06T3/40
CPCG06T3/4007
Inventor 钟宝江陆志芳
Owner SUZHOU UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap