1 INTRODUCTIONINTRODUCTION Multiresolution Meshes in Surface Modeling Leila De Floriani DISI - University of Genova (Italy) and UMIACS Joint work with Paola Magillo and Enrico Puppo (DISI - University of Genova)
2 INTRODUCTIONINTRODUCTION Outline o Introduction: motivations and background o LOD Models: layered versus multiresolution models o A general framework for multiresolution models: the Multi- Triangulation o Classification of Multiresolution Models: tree-like and evolutionary models o Tree-like Models: nested models based on regular or irregular subdivisions o Evolutionary Models: construction strategies and encoding structures o A taxonomy for multiresolution models
3 INTRODUCTIONINTRODUCTION Motivations o High complexity of 3D scenes mautomatic acquisition of the surface of solid objects Grange scanners -- 3D scanners Gresolution supported: ~ 10 facets / mm 2 mstandard solid modeling tools (CAD) Gcomplex 3D object defined by # faces > 100K mdigital terrain models Gmillions of faces mtesselation of implicit surfaces
4 INTRODUCTIONINTRODUCTION...Motivations... o 3D graphics is a limited, valuable resource mgraphics throughput of low level ws / pc : ~100K faces/sec minteractivity requires multiple frames per second mweb graphics (VRML) needs trasmission of data on low bandwidth networks Gscientific users: KB/sec local, KB/sec remote Gcommercial/home users: 56Kb/sec (in Italy) graphics file size: G24 byte/vertex ( if binary, >> if ascii )
5 INTRODUCTIONINTRODUCTION Reducing Graphics Costs o Visualization Stage: mculling back faces mview frustum culling mvisibility culling o Modeling Stage: mtessellate surfaces with triangle meshes msimplify meshes mconstruct a LOD model
6 INTRODUCTIONINTRODUCTION View Frustum Culling o an example of view frustum culling (images by SGI, OpenGL Optimizer) view frustum culling OFFculling ON
7 INTRODUCTIONINTRODUCTION Visibility Culling o an example of occlusion/visibility culling (images by SGI, OpenGL Optimizer)
8 INTRODUCTIONINTRODUCTION Level of Detail: Approximating surfaces with triangle meshes o Assumption: maccuracy of the approximation is proportional to the number of triangles o Objective: malways produce the simplest mesh that satisfies the accuracy required by the application
9 INTRODUCTIONINTRODUCTION …Approximating Surfaces... o On-the-fly simplification: extract from raw data a mesh of minimal size whose accuracy is sufficient for application needs Gonly raw data and the simplified mesh are stored Gsimplification is usually an expensive task
10 INTRODUCTIONINTRODUCTION …Approximating Surfaces... o LOD / multiresolution model: build a model off-line that encompasses many different representations and that can be queried efficiently Gmore expensive in terms of space Gmore efficient: support to real-time operations LOD model