Title: Euler’s elastica based models and algorithms in image processing

Abstract: Elastica regularizers have been studied in recent years by the image processing community. In this talk, we will first discuss and compare properties of segmentation models with  L^1 and L^2 Elastica regularizers by using a  novel augmented Lagrangian method based algorithm that captures curvature faithfully. We will then discuss a new denoising model with the L^1 Euler’s elastica as the regularizer, and also a novel ADMM based algorithm for this model with every subproblem being solved in a closed form.