Related Experiment Video
Updated: May 17, 2026
07:50
Measuring Local Tissue Strains in Tendons via Open-Source Digital Image Correlation
Published on: January 27, 2023
Nonlocal image restoration with bilateral variance estimation: a low-rank approach
Weisheng Dong1, Guangming Shi, Xin Li
1Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education, School of Electronic Engineering, Xidian University, Xi’an 710071, China. wsdong@mail.xidian.edu.cn
Summary
This study introduces a low-rank approach to simultaneous sparse coding (SSC), offering a new interpretation for natural image representation. The developed Spatially Adaptive Iterative Singular-Value Thresholding (SAIST) algorithm improves image restoration tasks like denoising and completion.
Area of Science:
- Computer Vision
- Image Processing
- Signal Processing
Background:
- Simultaneous sparse coding (SSC) offers advanced nonlocal image representation for low-level vision tasks.
- Existing methods like BM3D and LSSC demonstrate SSC's potential, yet lack a clear physical explanation for its effectiveness on natural images.
- Sparsity optimization, particularly with dictionary learning, presents significant computational challenges.
Purpose of the Study:
- To provide a physically plausible interpretation of SSC for natural images using a low-rank approach.
- To develop a computationally efficient and conceptually simple image restoration algorithm inspired by bilateral variance estimation.
- To enhance image denoising and completion techniques by generalizing local to nonlocal models.
Main Methods:
- A low-rank approach is applied to simultaneous sparse coding (SSC).
- Singular-value decomposition (SVD) of similar image patches is used for bilateral variance estimation, pooling local and nonlocal information.
- A novel algorithm, Spatially Adaptive Iterative Singular-Value Thresholding (SAIST), is developed for image restoration.
Main Results:
- SAIST generalizes BayesShrink to nonlocal models for noise reduction and extends sparsity optimization for image completion.
- The algorithm demonstrates conceptual simplicity and computational efficiency.
- SAIST achieves competitive or superior objective performance in image denoising and completion compared to state-of-the-art methods, especially under challenging conditions (high noise, significant missing data).
Conclusions:
- The proposed low-rank interpretation offers a new perspective on SSC for natural images.
- SAIST provides an effective and efficient solution for image restoration tasks, outperforming existing methods in key areas.
- The algorithm's performance favorably compares to current techniques in subjective quality assessments.
Related Concept Videos
01:24One-Compartment Open Model: Wagner-Nelson and Loo Riegelman Method for ka Estimation
This lesson introduces two critical methods in pharmacokinetics, the Wagner-Nelson and Loo-Riegelman methods, used for estimating the absorption rate constant (ka) for drugs administered via non-intravenous routes. The Wagner-Nelson method relates ka to the plasma concentration derived from the slope of a semilog percent unabsorbed time plot. However, it is limited to drugs with one-compartment kinetics and can be impacted by factors like gastrointestinal motility or enzymatic degradation.
On...
On...
01:11Residuals and Least-Squares Property
The vertical distance between the actual value of y and the estimated value of y. In other words, it measures the vertical distance between the actual data point and the predicted point on the line
If the observed data point lies above the line, the residual is positive, and the line underestimates the actual data value for y. If the observed data point lies below the line, the residual is negative, and the line overestimates the actual data value for y.
The process of fitting the best-fit...
If the observed data point lies above the line, the residual is positive, and the line underestimates the actual data value for y. If the observed data point lies below the line, the residual is negative, and the line overestimates the actual data value for y.
The process of fitting the best-fit...
01:26Linear Approximation in Frequency Domain
Linear systems are characterized by two main properties: superposition and homogeneity. Superposition allows the response to multiple inputs to be the sum of the responses to each individual input. Homogeneity ensures that scaling an input by a scalar results in the response being scaled by the same scalar.
In contrast, nonlinear systems do not inherently possess these properties. However, for small deviations around an operating point, a nonlinear system can often be approximated as linear.
In contrast, nonlinear systems do not inherently possess these properties. However, for small deviations around an operating point, a nonlinear system can often be approximated as linear.
01:29Application of Linearization and Approximation
A drone flying through complex terrain often relies on more than one sensing method to estimate small changes in altitude. Along with direct measurements, air pressure provides a useful indirect indicator of vertical movement. Atmospheric pressure decreases as altitude increases, and this relationship is commonly described using an exponential model. Although accurate, converting pressure measurements into altitude values requires calculations that are too complex to perform repeatedly during...
01:26Linearization and Approximation
Linearization is a mathematical technique used to approximate complex, nonlinear functions with simpler linear models in the vicinity of a chosen reference point. The method is based on the idea that, although a function may be difficult to evaluate exactly, its behavior near a specific input value can often be closely approximated by the tangent line at that point. This approach is particularly useful when small deviations from a known value are involved.Consider the square root function, for...