Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Low-rank tensor completion method for alpha-order total variation constraint of damaged video

A fully variable and broken technology, applied in image analysis, image data processing, instruments, etc., can solve problems such as blurred edges of structural fine-grained texture areas, repair images, loss of affine details, etc., and achieve the effect of overcoming the oscillation phenomenon

Pending Publication Date: 2021-06-18
XIAN UNIV OF TECH
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The existing LRTV is not enough to use the non-local and fine-grained structural information of tensors to repair images, which will lead to blurred edges and loss of affine details in areas with complex structures and fine-grained textures.

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
  • Low-rank tensor completion method for alpha-order total variation constraint of damaged video
  • Low-rank tensor completion method for alpha-order total variation constraint of damaged video
  • Low-rank tensor completion method for alpha-order total variation constraint of damaged video

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0137] The following will use YUV video data to illustrate the present invention's low-rank tensor completion method for the α-order total variation constraints of damaged video to further illustrate its effect:

[0138] The experimental data comes from YUV video sequences, and the video data are suzie and hall_qcif respectively. The experimental video data is read into MATLAB, some commonly used 4:2:0 YUV format video test sequences are used, and the first 100 frames are selected as the experimental data, so the data size is 176×144×100, they can be regarded as a 3D tensor. By randomly masking a part of the original tensor data in all channels of the experimental video data, the remaining pixels are used to form a damaged 3D tensor to complete the tensor Among them, the data loss rate of the experimental video is 95% and 75%. Simultaneously convert the three-dimensional tensor Expand along each module into a two-dimensional expansion matrix Here N=3, the sizes of the...

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 a low-rank tensor completion method for alpha-order total variation constraint of a damaged video, and the method is implemented according to the following steps: reading in a damaged video file with a high loss rate by utilizing MATLAB, processing the damaged video file into a three-dimensional tensor, expanding the three-dimensional tensor into a two-dimensional matrix along each module, performing regularization processing on boundaries of the matrixes, and obtaining a two-dimensional matrix of a zero Dirichlit boundary condition; defining a target functional about tensor completion of the damaged video, wherein the target functional comprises an alpha-order total variation regular constraint term and a low-rank constraint term, and the alpha-order total variation regular constraint term and the low-rank constraint term are not independent of each other; introducing three auxiliary matrixes for the boundary-regularized two-dimensional matrix, decoupling and optimizing a target functional through an augmented Lagrange formula, solving the optimized target functional, and finally obtaining a complemented three-dimensional video tensor through continuous iteration; and combining a fractional order TV regularization term in a fractional order bounded variation space with low-rank constraint to carry out tensor repair, so global information can be recovered, and lost fine details can also be recovered.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, and relates to a low-rank tensor completion method for α-order full variation constraints of damaged videos. Background technique [0002] Digital image inpainting refers to techniques that reconstruct completely missing / broken parts of an image or remove unwanted objects of interest in an imperceptible manner. In recent years, with the rapid development of data acquisition technology, a large number of multi-channel visual data sets have been collected in many areas of social production and life, such as: RGB images, digital videos, multi-spectral and hyperspectral images, etc. Among them, video data sets The scale and quantity of video are increasing day by day, especially in daily life, digital video occupies a very important position. However, due to the influence of transmission or compression, some information in video data is often lost or damaged. When the loss rate is h...

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
CPCG06T2207/10016G06T5/77Y02T10/40
Inventor 杨秀红薛怡许鹏石程金海燕
Owner XIAN UNIV OF TECH
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
Eureka Blog
Learn More
PatSnap group products