Alternating direction method of multipliers for nonlinear image restoration problems

Area of Science:

• Image Processing
• Computational Imaging
• Applied Mathematics

Background:

• Nonlinear image restoration is crucial for recovering degraded images.
• Common degradations include blur, system nonlinearity, and additive noise.
• Total Variation (TV) regularization is a powerful technique for image restoration.

Purpose of the Study:

• To develop an efficient numerical method for total variation (TV)-based nonlinear image restoration.
• To address image degradation from blur, system nonlinearity, and Gaussian white noise.
• To demonstrate the effectiveness of the proposed model and numerical scheme.

Main Methods:

• Formulation of an objective function with nonlinear least squares data-fitting and TV regularization terms.
• Development of an efficient alternating direction method of multipliers (ADMM) algorithm.
• Analysis of the convergence properties of the numerical scheme.

Main Results:

• The proposed alternating direction method of multipliers efficiently solves the nonlinear image restoration model.
• Numerical examples demonstrate the effectiveness of the model for nonlinear image restoration.
• The method shows promise for applications like high-dynamic range imaging.

Conclusions:

• The developed ADMM-based approach is effective for TV-based nonlinear image restoration.
• The numerical scheme is efficient and converges reliably.
• The proposed model and method offer a robust solution for complex image restoration tasks.