Visualizing Similarities between Volume Datasets by Using Critical Point Graph and Character Recognition Technique

Abstract

We propose a new technique for visually comparing the similarities between volume data sets by using Critical Point Graph (CPG) and Character Recognition Technique (CRT). Critical Points (CPs) represent feature points inside volume data and, CPG is a graph generated by connecting these points using streamlines based on vector field information. Passing Voxel Method (PVM) is a graph comparison technique which allows the visualization of both similarities and differences between two volume data. However, PVM has some limitations in its ability to judge similarities. Taking into consideration that streamlines of CPG are composed of small straight line segments, we have investigated the use of Directional Element Feature (DEF), usually used in CRT, by extending this technique for use in PVM in order to overcome some of its limitations.