Chunlin Wu, 南开大学


Augmented Lagrangian Method for Total Variation Related Problems over Triangulated Surfaces


Total variation regularization has been proven very useful in image processing and computer graphics applications. Recently many efforts have been contributed to efficiently solve this type of problems which are non-differentiable. Augmented Lagrangian method is one of the most efficient methods. In this talk, we will discuss this method for total variantion related problems over triangulated surfaces, including image denoising and segmentation on surfaces, as well as surface denoising.

« Back