This course will introduce the various techniques that enable the interactive visualization and handling of very complex objects and scenes. While there has been a leap in the power of graphics hardware, more complex datasets are generated through advances in 3D modeling, simulation and data capture. Thus, the need to deal with these massive models arises in fields such as scientific visualization, CAD, cultural heritage, videogame engines and others. The students will be exposed to hierarchical representation of scenes, model's simplification and visibility culling. As a result, they will gain a global view of the problem and a wide knowledge of the current solutions.
Weekly hours
Theory
2
Problems
0
Laboratory
1
Guided learning
0
Autonomous learning
5
Objectives
Using Hierarchical Geometric Models for the display of very large models.
Related competences:
CEE1.1,
CG3,
CTR5,
CTR6,
Simplification algorithms for triangle meshes.
Related competences:
CEE1.1,
CG3,
CTR5,
CTR6,
Visibility computation algorithms
Related competences:
CEE1.1,
CG3,
CTR5,
CTR6,
Interactive navigation in complex environments
Related competences:
CEE1.1,
CG3,
CB8,
CB9,
CTR5,
CTR6,
Contents
Hierarchical geometric models
Algorithms for space subdivision (regular grids, octrees, BSP trees, Kd-trees), scene subdivision (BVHs) and external memory-based data structures.
Mesh representation data structures
Triangle and polyognal mesh representation: Independent face set, Indexed face set, Adjacency lists, Winged edge, Half edge, Corner table.
Simplification of triangle meshes
Introduction to the basic concepts, operators and error metrics used in geometry and topology-based simplification. Its application to appearance-preserving simplification and out-of-core gigantic model simplification.
Level of detail
Introduction to object level of detail (LOD) and its application to complex scenes (time critical rendering). Strategies for LOD: Discrete, Continuous, or View-Dependent. Popping effect prevention.
Visibility computation
Introduction to the basic concepts and algorithms for visibility computation, including visibility preprocessing, point and region visibility, and visibility computation using the GPU. PVS compression.
Interactive navigation in complex environments
How to estructure gigantic data for out-of-core visualization of huge scenes. Use of view dependent visualization. Algorithms for collision detection in gigantic models.
Each student has to prepare the corresponding presentation and a supporting document, which have to sent to the course coordinator before the session. Objectives:1234 Week:
15
Theory
0h
Problems
0h
Laboratory
0h
Guided learning
3h
Autonomous learning
6h
Exercises
Set of exercises raised during the course to assess knowledge acquisition by students during the course.
Theory
0h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
9h
Teaching methodology
This course is structured in three session types:
* T sessions (theory): presentation by the corresponding professor. The professor will ask the students to do some short exercises on the subjects covered in these sessions.
* D sessions (discussion): sessions conducted by the professor, in which some students will solve exercises or present previously distributed papers. Each student has to prepare the corresponding presentation and a supporting document, which have to sent to the course coordinator before his D session.
* L sessions (lab): in these sessions students will have to solve practical problems programmings some of the algorithms presented in the theory sessions. L sessions will start with a short lecture section.