CMPT 464/764: Geometric Modelling in Computer Graphics (Fall 2024)

This course covers recent and advanced modeling techniques in computer graphics. Our focus will be on the acquisition, representation, processing, analysis, and synthesis of 3D shapes, with applications to real-time 3D graphics, generative AI, design and manufacturing, as well as 3D machine vision and robotics. We will cover a variety of 3D representations including polygonal meshes, point clouds, voxels, implicit functions, as well as structured representations. Both classical and modern-day learning-based approaches will be studied. This field is still fast evolving with many interesting problems and much aspiration for application development and future research, e.g., in geometric deep learning, computational design, and creative modeling. Basic mathematical concepts and tools necessary to understand the course will be presented depending on students background. Programming experience in computer graphics (CMPT 363 or equivalent) and machine learning (Python, C++) would be required. This course is cross-listed between CMPT 464 and 764.

Course Topics: (not necessarily covered in that order):

• Computer graphics and geometric modeling in the AI era
• Geometric modelling paradigms for 3D content creation
• 3D shape representations: parametric primitives, implicit functions, meshes (subdivision surfaces), point clouds, voxels, multi-view images, structured representations, and their "neuralization" (i.e., 3D shape representation learning)
• 3D shape acquisition and surface reconstruction
• Segmentation and correspondence for 3D vision
• 3D content creation and generative (including creative) modeling
• 3D printing

Grading:

One midterms (30%), one assignment (20%), and a final project (50%)

References:

• A Sampler of Useful Computational Tools for Applied Geometry, Computer Graphics, and Image Processing, Cohen-Or et al., 2015.

• Survey and research papers to be added throughout the course

Interests in the topic; no obvious symptoms of math phobia. Programming experience in computer graphics (CMPT 363 or equivalent) and machine learning (Python, C++) would be required.