CENG789 - Digital Geometry Processing
Objective: Digital Geometry Processing is a subfield of Computer Graphics. This introductive course teaches the geometry processing pipeline, which involves acquisition, analysis, manipulation, and printing. The practical applications implemented throughout the course content will make the student comfortable in engineering, medical, entertainment (VFX/games), and many other related disciplines that make heavily use of 3D mesh models.
Syllabus: As one file.
- Surface Representation
- Polygons and Triangulations
- Point Sets
- Mesh Data Structures and 3D Graphics Programming
- Distances and Sampling on Meshes
- Descriptors on Meshes
- Mesh Smoothing and Decimation
- Mesh Repairing and Tetrahedral Meshes
- Surface Reconstruction
- Mesh Parameterization
- Mesh Registration (Rigid-body)
- Mesh Deformation
- Data-driven Shape Analysis
- 3D Printing
- 40%: 2 Programming Assignments
- 40%: Term Project
- 20%: Final Exam
Time & Place: Tue 14:40-17:30 & A101
Textbook: None. Lecture notes will be sufficient. Use "Polygon Mesh Processing" by Botsch et al. as a reference.
Prerequisites : C/C++ programming required. Some background in Computer Graphics is helpful, but not necessary.
Instructor: Yusuf Sahilliošlu. Office: B107. Come in anytime when I'm in the office. Email: firstname.lastname@example.org
Please read this document
first. Keyword-based suggestions in that document are detailed below.
- Cage: Tight caging of a mesh.
- Correspondence: Map computation over multiple and/or noisy meshes.
- CSG: Constructive solid geometry.
- Deformation: Facial performance capture, a real-time registration task. See also this.
- Deformation: Character posing.
- Descriptor: Shape representatives.
- Fabrication: 3D printing.
- Parameterization: Mapping of a mesh in 3D into 2D.
- PCA: Synthesis of new shapes using the existing population.
- Reconstruction: Isosurfaces.
- Reconstruction: Geographic Information System (GIS), LiDAR.
- Repairing: Normal orientation correction.
- Retrieval: Bringing similar shapes from a database.
- Segmentation: Clustering meaningful parts.
- Simplification: Decreasing resolution while preserving details.
- Skeleton: Extracting the medial axis of a given mesh. Can also be rigged.
- Subdivision: Increasing resolution in a consistent and smooth manner.
- Surfacing: Hole filling.
- Tetrahedralization: Representing volume as a mesh.
- Interpolation/morphing: Topology-varying or data-driven.
Back to the home page