PaLEnTIR: A Parametric Level Set-based Approach to Image Reconstruction and Restoration

Explore an in-depth seminar presentation where Professor Misha Kilmer from Tufts University introduces PaLEnTIR, an innovative parametric level set-based approach for image reconstruction and restoration. Delve into the challenges of inverse problems across science and engineering applications, focusing on identifying regions of interest (ROIs) such as tumors in X-ray CT, subsurface contaminants in hydrological data, and buried objects in electromagnetic data. Learn about an alternative approach to traditional image reconstruction methods that uses a parameterized image model for piecewise constant objects. Discover how the PaLEnTIR model optimizes a small set of parameters with built-in regularization, eliminating the need to know the number of contrasts or their values in a given scene. Examine the model's properties, inversion algorithm, and its practical applications through demonstrations on 2D and 3D linear and non-linear inverse problems.