The Computer Graphics Group give students the opportunity to work on a wide variety of topics in the fields of Computer Graphics and Geometry Processing. Bachelors and Master students interested in those fields are encouraged to contact us to discuss theses topics. You can also find already-existing thesis topics below. All theses' topics are flexible and can be arranged to suit your interests.
Potential Theses Topics
"Interactive Volumetric Mesh Generation"
- Level: Bachelor or Master
- Topic description:
Volumetric meshes are often generated from closed 3D surfaces, with a wide variety of automatic techniques.
These techniques mainly focus on element quality or closeness to features, which are generally important requirements for the usage of volumetric meshes.
However, complex element configurations are often found inside automatically-generated volumetric meshes, which can make using mesh-processing methods (e.g. deformations or parametrizations) a very challenging task.
This project thus aims at designing and developing a system that will allow to take a reverse approach, by interactively generating volumetric meshes from scratch.
By explicitly generating complex and/or problematic structures, one will be able to challenge exisiting and future mesh-processing methods in a more focused way.
A good starting point will be to design a set of "tools" that will minimize the manual effort to create complex meshes and configurations.
Tools that are generally easy to describe verbally, but difficult to define formally.
- Goal:
The main goal of this project is to create an interactive system to interactively and efficiently generate volumetric meshes based on C++. The student will have to design dedicated tools for this task, as well as implementing and testing them.
- Requirements:
- Good C++ skills.
- Interest in meshing, toplogy.
- Interest in interactions and interfacing.
- Interest in open-ended, pure research.
"Interactive Volumetric Mesh Generation"
- Level: Bachelor or Master
- Topic description:
Volumetric meshes are often generated from closed 3D surfaces, with a wide variety of automatic techniques.
These techniques mainly focus on element quality or closeness to features, which are generally important requirements for the usage of volumetric meshes.
However, complex element configurations are often found inside automatically-generated volumetric meshes, which can make using mesh-processing methods (e.g. deformations or parametrizations) a very challenging task.
This project thus aims at designing and developing a system that will allow to take a reverse approach, by interactively generating volumetric meshes from scratch. By explicitly generating complex and/or problematic structures, one will be able to challenge exisiting and future mesh-processing methods in a more focused way.
A good starting point will be to design a set of "tools" that will minimize the manual effort to create complex meshes and configurations. Tools that are generally easy to describe verbally, but difficult to define formally. - Goal: The main goal of this project is to create an interactive system to interactively and efficiently generate volumetric meshes based on C++. The student will have to design dedicated tools for this task, as well as implementing and testing them.
- Requirements:
- Good C++ skills.
- Interest in meshing, toplogy.
- Interest in interactions and interfacing.
- Interest in open-ended, pure research.
"Volumetric Mesh Smart Visualization"
- Level: Bachelor or Master
- Topic description: While working on volumetric meshes, researchers often need to study specific areas in a direct and visual manner. Those Regions of Interest (ROI) can, for instance, be problematic situations were some algorithm fails or present singular and interesting topological/connectivity properties.
However, visualizing ROIs can be quite challenging when they are located in the interior of meshes, simply because closer elements of the mesh obstruct our field of view. We are thus looking for novel ways to make the visual analysis of such ROIs easier and more efficient. Early ideas include but are not limited to smart transparency filters, camera focus, local animation, etc.
Working on this project will include a mix of design, experimentation and C++ development. It is a rather open project, with lots of room for students to be creative in their design.
- Goal: Designing visualization techniques that allow for better visual analysis of Regions of Interest. Then implementing those tools, experiment with them and re-iterate to improve them.
- Requirements:
- Good C++ skills.
- Interest in visualization and interaction development.
- Interest in open-ended, pure research.
Finished Master Thesis:
"3D building planimetry from unaligned point clouds" by Dave Meier. July 2022. report | code
Finished Bachelor Thesis:
"3D Metric Fields - Optimizing Frame Fields in a new Metric" by Florin Achermann. October 2023. report
"Fast Hexahedral Mesh Extraction from Locally Injective Integer-Grid Maps" by Tobias Kohler. October 2023. report | slides
"Up to 58 Tets/Hex to untangle Hex meshes" by Luca Schaller. November 2022. report | code | slides
"Rust-V Physically Based Spectral Rendering" by Julius Oeftiger. July 2022. report | code
"Creation and modification of 3D meshes in virtual reality" by Marcel Zauder. February 2022
"Foldover-Free Maps: an Evaluation" by Maksim Fomin. January 2022. report
"Hex-Mesh Optimization with Edge-Cone" by Elias Wipfli. December 2021
"Geodesic distance computation via virtual source propagation in a volumetric setting" by Marco Cacciatore. October 2021.
"Optimized visualization of OpenVolumeMesh in Blender" by Lorenzo Wipfli. October 2021.
"Topology preserving Meshing of Medical Data with multiple Materials" by Steve Mürset. September 2021.
"Feature Detection in Triangle Meshes" by Lukas Seeholzer. March 2021. report
"Flesh Simulation with Application to Character Animation" by Corina Danja Masanti. January 2021. report
