Interactive Illustrative Visualization of 3D Models
Abstract
In this thesis I deal with three problems related to interactive illustrative visualization of 3d models.
The first problem is visualization of interior structures. I present a novel method for visualization of interior structures. The method is capable to reveal all interior structures at once or to visually emphasize only several important 3d objects. The occlusion of the 3d objects is eliminated with opacity modulation. I combine several properties of the objects into the opacity of the 3d objects. To do so I use fuzzy sets, fuzzy logic and multi-criteria decision making.
The second problem is visualization of functional relations. I present a novel method for visualization of functional relations of selected part of the 3d model. The occlusion of the 3d objects is eliminated with method described above. The partonomic organization of the 3d model and contacts between individual 3d objects are utilized to estimate the functional relations. The main benefit of the method is in using the partonomic organization and contacts between the 3d objects as that they are domain independent. The results show that the method provides aesthetic visualization of functional relations.
The third problem is interconnection of verbal and visual information through external labeling. I present a novel method for computing the labeling of 3d illustrations in real-time. I show how to reduce the dimensionality of the problem and how to formulate the labeling problem as a multiple criteria optimization problem. I solve the optimization problem using fuzzy logic combined with greedy optimization. The results show that the method compares favorably to the state-of-the-art techniques for interactive external labeling.
All three presented methods achieve real-time performance and thus are suitable for generation of both static a interactive illustrations.
Download
All techniques described in the thesis were implemented with Tiger library.