Logo Logo
Switch Language to German
Kolb, Daniel ORCID: 0000-0001-9645-3319 (December 2016): Mesh based Scene Evaluation Metrics for LOD and Simplification. Master Thesis, Faculty of Mathematics, Computer Science and Statistics, Ludwig-Maximilians-Universität München.


I present seven metrics to quantify attributes of different meshes in a scene. Each metricrepresents a different geometrical or topological aspect of the mesh. Theresulting ratingvalues serve to convey the underlying complex data to the user. These allow the user toswiftly compare several features of multiple meshes. The metricsmay thus guide usersand programs during the process of mesh modification, i.e. optimization, simplification orsmoothing, and scene modification as a whole.I evaluate each metric individually by applying them to a sample scene. To examine thecorrectness and expressiveness of the metrics I compare the automatically calculated ratingsto the raw base data. I find two of the metrics to be immediately useful and four of theratings promising, but in need of adjustments. The remaining last metric, however, requiressignificant rework to generate useful data on par with the other six metrics.This thesis first introduces the subject with a motivating example. It then presents importantconcepts and research on related topics. Afterwards it details the concept of the programand the mathematical considerations it is based on. It also lists my approach to solving thechallenges which emerged during the implementation. Subsequently, the thesis focusses onthe visualized output of the program and challenges said ouput. Finally, it contrasts theexpectations and goals of each metric with the respective actual results