Real-time advanced visualization techniques for the development of applications in the cultural heritage field

Bachelor Thesis

Let’s consider a statue made of marble. If we decided to scan it digitally to obtain a 3D representation on a computer, we would end up with a very detailed model, albeit void of any information regarding its material. Visualizing this extremely detailed 3D model would result in a very flat, uninteresting and fake image.

To partially address this problem, we could apply on it some pictures of some kind of marble which resembles the one used by the sculptor to create the statue.

Doing this alone won’t give it justice, as we’ve just added the colour information but we’re still missing a fundamental aspect: how it interacts with the ambient light.

If we imagine the statue at the centre of a sphere where, on its surface and pointing inward, we would place a sufficiently high number of cameras, we could obtain a pretty accurate approximation of how shadowy or lit each small section of the statue’s surface is as it’s illuminated by the environment.

But light doesn’t just bounce on the surface; illumination of a surface doesn’t just come from the outside, where the light directly hits the surface, but also from the inside. The light is capable of penetrating surfaces, bouncing inside most materials and the layers they’re characterized by, and then exiting from another pseudo-random location.

In my thesis I present 3 techniques aimed at solving the above problems in a constrained environment: real-time computer visualization, where the main challenge is to make sure that wathever algorithm we decided to implement, it should complete in less than 30ms.

Download

You can download my thesis from this link (size: 14MB). Hope you enjoy it!