CSE - Procedural Level-of-Detail and Dynamics in Tangent Spaces

Under the constraints of real-time performance, high-fidelity computer graphics require heuristics and simplifications to replicate complex phenomena. This commonly involves working in a two-dimensional (2D) surface despite, for instance, representing graphics in a three-dimensional (3D) space. This project involves using such a 2D space to simulate surface gradient changes of a volume of uniform particles when interacted with by external forces, i.e., given a volume of consistent particles, how does its surface change shape when interacted with, based on the attributes of the particles?

This will require determining an appropriate number of variables for parametrization to ensure granular surfaces (e.g., from sand to snow) are able to be simulated with some amount of accuracy. This project is an extension of contemporary techniques which only provide one dimension of change across a surface, whereas we aim to achieve a multivariate approach to more accurately simulate particle displacement in 2D textures.

Separately, this project requires software validation and ideally an analysis of metrics (including their faults) when releasing interactive software in failing, partial, or complete states. In particular, this review is for interactive computer graphics applications, like video games. Understanding appropriate evaluation techniques, and assessing any potential gaps, is meaningful to help assess the state of readiness for software releases in an industry which currently relies on mass testing post release for validation.