Prof. Dr. Vladimir Molchanov (Leibniz-Fachhochschule Hannover)
Title: Geometry of Perceptual Color Space
anschl. After-Lecture-Beer
Hier anmelden!

Abstract:
Color plays a fundamental role in how we understand and communicate the world around us: from scientific visualization and digital media to art and everyday perception. Yet, color is surprisingly difficult to describe mathematically. Understanding how we perceive color brings together biology, psychology, computer science, mathematics and other fields.

One way to study this problem is to think of perceived colors as points in a geometric space, where the distance between two points represents how different the colors appear to us. This idea goes back to Riemann, who famously used the system of colors as an example of a multidimensional manifold in his 1854 habilitation lecture on the foundations of geometry. Later, Helmholtz introduced a Riemannian geometry of color, and Schrödinger further developed these ideas.

But how well does this geometric picture actually describe human color perception? Recent work by Bujack et al. has shown that the principle of diminishing returns, where a large color difference is perceived as smaller than the sum of several smaller differences, cannot be explained by a global Riemannian model. This raises a natural question: if the Riemannian model does not describe color space globally, how well does it describe it locally? In this talk, I explore this question by investigating how the geometry of perceived color varies across color space and with direction.