I Can See Clearer Now: The Optimal Regularization Parameter - Imaging and Inverse Problems Seminar

Explore advanced techniques for image restoration in this virtual seminar from the 31st Imaging & Inverse Problems (IMAGINE) OneWorld SIAM-IS series. Delve into the crucial topic of selecting optimal regularization parameters in variational frameworks for image restoration. Learn about traditional methods and discover a fully automatic approach for parameter selection in cases of space-invariant blur and additive white Gaussian noise. Understand how the residual whiteness principle can be applied to various variational models, including those with Tikhonov and Total Variation regularizers. Gain insights into implementing this strategy for non-quadratic regularizers using an iterative optimization scheme based on the alternating direction method of multipliers. Discover how this approach can be extended to tackle the challenging inverse problem of image super-resolution.