Convergence and Non-Convergence of Algebraic Iterative Reconstruction Methods - Imaging and Inverse Problems Seminar

Explore the convergence and non-convergence of algebraic iterative reconstruction methods in this virtual seminar talk from the 34th Imaging & Inverse Problems (IMAGINE) OneWorld SIAM-IS series. Delve into the intricacies of limited-angle and limited-data X-ray CT problems, focusing on methods like ART (Kaczmarz), SART, and SIRT. Examine the consequences of using unmatched pairs in forward and back projections, including potential accuracy deterioration and convergence failure. Discover novel approaches to address non-convergence issues, such as modifying iterative methods or implementing AB- and BA-GMRES techniques. Gain insights from speaker Per Christian Hansen of DTU, who presents recent theoretical and computational findings in this collaborative research effort.