We implement a framework for view-dependent refinement of progressive meshes. We use a similar approach to Hoppe’s progressive mesh (PM) representation (ACM Comput. Graphics, Proceedings of SIGGRAPH’97, August 1997, pp. 189–198) for view-dependent refinement with a different algorithm for constructing PM representation. Our method is simple to implement and fast enough to achieve interactive frame rates for moderately complex models (models containing hundreds of thousands of polygons) on a machine with polygon rendering hardware. Moreover, our implementation allows changes to topology and achieves a simpler and sometimes more realistic refinements. We use the original mesh and perform edge collapses so that the detail level is adjusted to the desired level. Detail level is dictated by the view frustum, surface orientation and screen space error. For example, the portions of the mesh outside the viewing frustum or the parts that look away from you are refined to a lower detail level. The refinement step also takes the screen space error into account, i.e.: distant parts are fined more, flat areas refined more whereas silhouettes are refined less. Our implementation allows changes to topology and achieves a simpler and sometimes more realistic refinements. For example, during the refinement, unconnected parts of the mesh can merge together (this depends on the underlying edge decimator. We used Garland's Simplification by Quadric Error Metrics). Below, you can see some animated view-dependent refinement examples. Green lines describes the viewing frustum according to a phantom camera, to demonstrate the view dependent refinement process. If the user does not spawn a phantom camera, the refinement process is done with respect to the normal camera whose image plane is the whole window.
We developed a framework for the stereoscopic view-dependent visualization of large scale terrain models. We use a quadtree-based multiresolution representation for the terrain data. This structure is queried to obtain the view-dependent approximations of the terrain model at different levels of detail. In order not to lose depth information, which is crucial for the stereoscopic visualization, we make use of a different simplification criterion, namely, distance-based angular error threshold. We also present an algorithm for the construction of stereo pairs in order to speed up the view-dependent stereoscopic visualization. The approach we use is the simultaneous generation of the triangles for two stereo images using a single draw-list so that the view frustum culling and vertex activation is done only once for each frame. The cracking problem is solved using the dependency information stored for each vertex. We eliminate the popping artifacts that can occur while switching between different resolutions of the data using morphing.
February 1999 - February 2001