Towards corner matching for image reconstruction

摘要

A common problem in so-called multi-source visualization data analysis and211u001evisualization is the identification of certain common features in two-dimensional 211u001e(2D) images and their counterparts in three-dimensions (3D). We discuss methods 211u001eto define and effectively extract features, e.g., corners or edges, from 2D 211u001eimages and 3D models. Work toward this goal is described and lessons learned 211u001ediscussed.