Image denoising method combining Tetrolet transform domain and PDE (Partial Differential Equation) and GCV (Generalized Cross Validation) theory

A technique of transforming domains and images, applied in the field of image denoising based on Tetrolet transform, which can solve the problems of image block effects and visual effects that need to be improved

Inactive Publication Date: 2014-03-26
ZHEJIANG NORMAL UNIVERSITY
View PDF2 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The disadvantage is that for images with rich image texture and details, there is a

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 denoising method combining Tetrolet transform domain and PDE (Partial Differential Equation) and GCV (Generalized Cross Validation) theory
  • Image denoising method combining Tetrolet transform domain and PDE (Partial Differential Equation) and GCV (Generalized Cross Validation) theory
  • Image denoising method combining Tetrolet transform domain and PDE (Partial Differential Equation) and GCV (Generalized Cross Validation) theory

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0063] image 3 (a) is a standard Lena (512×512) experimental test image, the denoising results of various methods when adding σ=20 Gaussian white noise are as follows Figure 4 Shown. From Figure 4 It can be seen that various denoising images combined with PM1 methods have the best effect. Figure 4 (c) The denoising image of Wavelet transform combined with PM1 method has grid-like linear stripes, especially on Lena's face; the denoising image of Contourlet transform combined with PM1 is more blurred than the denoising image of Wavelet transform combined with PDE. However, it can be seen from the brim that the edge processing has been strengthened, relatively smooth, and there is no jagged phenomenon; Curvelet transformation combined with PM1 denoising image eliminates the current grid situation generated by the above two transformations combined with PDE method, but Lena face There are still a little blocky spots; the denoised image of Shearlet transform combined with PM1 has...

Embodiment 2

[0071] image 3 (b) is a geostationary satellite cloud image (256×256) image, the denoising results of various methods when adding σ=25 Gaussian white noise are as follows Image 6 Shown. From Image 6 It can be seen that various denoising images combined with PM1 methods have the best effect. Image 6 (c) In the wavelet transform combined with PM1 method, the denoising image is blurred, and the boundary of the typhoon eye is not clear and has stripes; the denoising image of Contourlet transform combined with PM1 also has linear stripes, especially in dark areas, and the boundary is not very large. Clear; the denoising image of Curvelet transform combined with PM1 is blurry, especially the dark gray background part and the large white cloud part is not clear, and there is block noise; the denoising image of Shearlet transform combined with PM1 is compared with the previous methods, denoising The effect is good, but there are fine horizontal stripes; the denoising image of the Te...

Embodiment 3

[0079] Due to the above experiments, the denoising algorithm in this paper and the Tetrolet transform combined with the PDE denoising algorithm both use the pair Δu=u 0 -u c Perform PDE processing, and then perform multiple iterations to reach the iteration termination condition. In the above experiment, we set the iteration termination condition to be 10 times. At the same time, the two methods have better denoising effects, so we will discuss how to change the iteration termination condition. The effect of noise results. Since the above experiment has shown that the denoising result of the combined PM1 method is better than the denoising result of the PM2 and TV methods, only the denoising method combined with the PM1 method is used here. Select Lena noise-added image (σ=20) as the experimental object, use the "T_GCV_PM1" denoising method and "T_PM1" denoising method proposed in this article, and set the iteration termination conditions as 1, 3, 5, 10, 15, 20, 30 times for co...

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 aims to provide an image denoising method based on Tetrolet transform. In the method, a partial differential equation (PDE) and generalized cross validation (PDE) are combined, thereby achieving a better noise denoising effect. The method comprises the following steps: performing Tetrolet transform on a noise-including image; determining an optimal denoising threshold by using a GCV theory in a transform domain and performing threshold processing; performing Tetrolet inverse transform to obtain an initial denoised image; smoothening the image by using a PDE denoising model under the consideration that a square block effect is generated by the Tetrolet transform in order to keep the edge information of the image. The PDE comprises three denoising models, namely, PM1, PM2 and TV. Meanwhile, compared with other five kinds of denoising results based on a multi-scale transform method, the method has the advantages of high PSNR (Peak Signal Noise Ratio) value of the denoising result, good image vision effect, outstanding details, further improvement on the subjective and objective quality of the denoised image and great promotion of subsequent work of the denoised image.

Description

Technical field [0001] The invention belongs to the field of image denoising. Specifically, it relates to an image denoising method based on Tetrolet transform for the purpose of improving image quality due to noise interference. Background technique [0002] In the process of video image acquisition, processing and transmission, it is often affected by the interference of imaging equipment and external environmental noise, resulting in the degradation of image quality. With the continuous improvement of wavelet analysis theory, wavelet analysis has been widely used in image denoising. The existing wavelet denoising methods mainly include: nonlinear wavelet transform threshold denoising method, tree wavelet denoising method, multiwavelet denoising method and wavelet coefficient model denoising method. Among them, Florian Luisier et al. proposed a fast algorithm based on Haar wavelet transform to eliminate Poisson noise in images, and its non-redundant inter-scale wavelet thresh...

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
IPC IPC(8): G06T5/00
Inventor 张长江陈源
Owner ZHEJIANG NORMAL UNIVERSITY
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