4:30pm -5:00pm

Dr. Iván Ojeda-Ruiz

Mathematics, Texas State University

Title: A Fast Normalized Cut with Automated Constraints via Graph Coarsening

Abstract: We demonstrate the K-way data clustering algorithm where constraints are selected automatically from the data by using a multiscale coarsening algorithm. Our algorithm is formulated in a much more effective way than the standard Rayleigh-Ritz projection algorithm for image segmentation with a priori partial grouping constraints. Additionally, unlike manual choices of constraints, our constraints are selected automatically by using the multiscale graph coarsening algorithm. The graph coarsening algorithm adopted from the so-called Segmentation by Weighted Aggregation is capable of performing image segmentation by reconstructing the image based on the coarsened graph using the prolongation operator. Unlike their algorithm, we segment the image in the finest scale by using the K-way Constrained Normalized Cut with such coarsened graph as constraints. We present a number of segmentation results that can demonstrate the superiority of our algorithm. The work is a joint project with Dr. Young Ju Lee.