I'll discuss some interesting problem statements arising from ongoing computer graphics, machine vision, and computational geometry projects. Some of these involve novel input or cost models. Some involve sheer complexities that dictate certain limits on combinatorics. Some involve hard real-time constraints arising from (e.g.) rendering. All have arisen in practice. To my knowledge, few are well-addressed by current threads of geometric research.I'll give examples from (at least) rendering, global illumination, conservative visibility, ray-casting, construction of arrangments, automatic machine vision, and approximation within poly-time algorithms.
This talk will have partial, but not total, overlap with one given at Dagstuhl in February.