392153 Visual Computing (V) (SoSe 2025)

This course unites topics in computer graphics, computer vision, and human perception---visual computing.
It has a strong practical component with students building the foundations of Gaussian splatting---the state-of-the-art algorithm for 3D reconstruction and rendering.

Conceptually, the lectures introduce computer graphics from a human-centric point of view. Human perception motivates the design of extended reality devices and efficient computer graphics algorithms. How can we simulate how light propagates in real time by exploiting the limitations of the human visual system? In turn, the computer graphics models are used for 3D computer vision by inverse graphics, the analysis by synthesis principle. How can we invert the image, which measures light intensity?

The basic building blocks for visual computing applications that are taught are:

• rendering techniques for computer graphics (ray-tracing, splatting, rasterization),
• representations for objects (sum of Gaussians)
• 3D geometry (perspective projection, optimization)
• human (color) perception models

Together, they form the foundation for developing and extending Gaussian Splatting. Practical experiences are gained in the assignments and tutorials.

The following slides provide additional details and context to the existing visual computing modules at Bielefeld University:
https://docs.google.com/presentation/d/e/2PACX-1vRGyE0j_cU324qp-vOEE4hEBPn3HdwliZQqeg_vT0w7Yt9EUXmgFB6NamL_ynqeBMYRgD-cdvfvXobA/pub?start=false&loop=false&delayms=3000

Requirements for participation, required level

The course builds upon the programming skills, mathematical models, and concepts learned in the introductory Programming and Math courses.
Required are basic programming skills (in python or Julia) and basic mathematics (in particular linear algebra, some integral calculus, little probability theory).

Prior knowledge of the PyTorch framework typically used for machine learning is a plus but not required.

Bibliography

The course material is largely self-contained and refers to external sources where not. The https://scratchapixel.com/ website is an excellent resource for self study of the covered materials and beyond. A script covering the core methodologies may be provided in future iterations of this course.

External comments page

https://docs.google.com/presentation/d/e/2PACX-1vRGyE0j_cU324qp-vOEE4hEBPn3HdwliZQqeg_vT0w7Yt9EUXmgFB6NamL_ynqeBMYRgD-cdvfvXobA/pub?start=false&loop=false&delayms=3000