Unlock instant, AI-driven research and patent intelligence for your innovation.

Anisotropic Gaussian side window kernel constrained image guided filtering method

An anisotropic, image-guided technology, applied in the field of image processing, can solve problems such as limited rectangular window width and small edge resolution, and achieve the effects of reduced complexity, clear edges, and simple filtering methods

Active Publication Date: 2021-01-22
XIAN UNIV OF POSTS & TELECOMM
View PDF14 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in order to achieve the ideal edge preservation effect, the side window filtering method will be limited by the width of the rectangular window. The larger the width, the smaller the edge resolution.

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
  • Anisotropic Gaussian side window kernel constrained image guided filtering method
  • Anisotropic Gaussian side window kernel constrained image guided filtering method
  • Anisotropic Gaussian side window kernel constrained image guided filtering method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0065] exist figure 1 Among them, the image-guided filtering method constrained by the anisotropic Gaussian side window kernel of the present embodiment consists of the following steps:

[0066] (1) Construct an anisotropic Gaussian kernel

[0067] Construct the anisotropic Gaussian kernel g as follows σ,ρ,θ (n):

[0068]

[0069]

[0070] Where n is the local pixel position in the filter window, θ is the rotation angle based on the y-axis, θ∈(0,π], σ is the Gaussian scale, σ∈(1,6], the value of σ in this embodiment is 3, ρ is the anisotropy factor, ρ∈(1,12], the value of ρ in this embodiment is 6, R θ is the rotation matrix with direction θ.

[0071] (2) Determine the anisotropic Gaussian side window kernel

[0072] Determine the anisotropic Gaussian side window kernel N according to formula (3) θ :

[0073] N θ ={n|xcosθ+ysinθ>0,g σ,ρ,θ (n)>ε,n=[x,y]} (3)

[0074] Where x and y are non-negative integers, ε is a threshold, ε∈[0.00005,0.00015], and the value of...

Embodiment 2

[0117] The image-guided filtering method constrained by the anisotropic Gaussian side window kernel of the present embodiment consists of the following steps:

[0118] (1) Construct an anisotropic Gaussian kernel

[0119] Construct the anisotropic Gaussian kernel g as follows σ,ρ,θ (n):

[0120]

[0121]

[0122] Where n is the local pixel position in the filter window, θ is the rotation angle based on the y-axis, θ∈(0,π], σ is the Gaussian scale, σ∈(1,6], the value of σ in this embodiment is 1.5, ρ is the anisotropy factor, ρ∈(1,12], the value of ρ in this embodiment is 1.5, R θ is the rotation matrix with direction θ.

[0123] (2) Determine the anisotropic Gaussian side window kernel

[0124] Determine the anisotropic Gaussian side window kernel N according to formula (3) θ :

[0125] N θ ={n|xcosθ+ysinθ>0,g σ,ρ,θ (n)>ε,n=[x,y]} (3)

[0126] Where x and y are non-negative integers, ε is a threshold, ε∈[0.00005,0.00015], and the value of ε in this embodiment is...

Embodiment 3

[0129] The image-guided filtering method constrained by the anisotropic Gaussian side window kernel of the present embodiment consists of the following steps:

[0130] (1) Construct an anisotropic Gaussian kernel

[0131] Construct the anisotropic Gaussian kernel g as follows σ,ρ,θ (n):

[0132]

[0133]

[0134] Where n is the local pixel position in the filter window, θ is the rotation angle based on the y-axis, θ∈(0,π], σ is the Gaussian scale, σ∈(1,6], the value of σ in this embodiment is 6, ρ is the anisotropy factor, ρ∈(1,12], the value of ρ in this embodiment is 12, R θ is the rotation matrix with direction θ.

[0135] (2) Determine the anisotropic Gaussian side window kernel

[0136] Determine the anisotropic Gaussian side window kernel N according to formula (3) θ :

[0137] N θ ={n|xcosθ+ysinθ>0,g σ,ρ,θ (n)>ε,n=[x,y]} (3)

[0138] Where x and y are non-negative integers, ε is a threshold, ε∈[0.00005,0.00015], and the value of ε in this embodiment is 0.00...

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 anisotropic Gaussian side window kernel constrained image guided filtering method. The method comprises seven steps of constructing an anisotropic Gaussian kernel, determining an anisotropic Gaussian side window kernel, determining the radius of the anisotropic Gaussian side window kernel, determining an anisotropic Gaussian side window kernel weight matrix, constructingan anisotropic Gaussian side window kernel guided filter, selecting an optimal filtering result and determining a filtering result. Due to the fact that the anisotropic Gaussian edge window kernel isadopted, compared with a traditional box filtering method, the method has the advantages that the edge of the image processed through the method is clear, and the retentivity is good while the strongedge preserving effect is achieved. When guided filtering processing is carried out, weighting filtering is carried out by adopting the weight coefficient of anisotropic Gaussian side window kernel filtering, and better edge retentivity is achieved. And a linear filtering method is adopted, so that the complexity of the filtering method is greatly reduced. The method has the advantages of being simple in filtering method, convenient to operate, good in high-resolution edge retentivity and the like, and can be used for image filtering processing.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to image edge processing. [0002] technical background [0003] Edge-preserving filtering methods are often used in preprocessing operations of computer vision and graphics and image processing, and the quality of the results directly affects many subsequent operations. The edge-preserving filtering method not only pays attention to the smoothing of the image, but also pays attention to maintaining the edge details. Traditional image smoothing methods focus on the smoothing effect, resulting in the loss of edge details after the image is filtered. In order to solve the technical problem of loss of edge details, many edge-preserving filtering methods have been proposed and widely used in computer animation, digital photography and other technical fields. [0004] Image filtering methods based on local filters use adjacent pixels to calculate new pixels, such as...

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): G06T5/00
CPCG06T2207/20192G06T5/70
Inventor 王富平吉聪聪陈鹏博公衍超高梓铭刘颖韦同胜刘卫华李兴
Owner XIAN UNIV OF POSTS & TELECOMM