Procedural Methods for Images

Aim

In order to complete the course, the student must be able to:

• describe images and geometries as tools in modern computer graphics.
• describe classic and modern methods for procedural generation of images, especially textures that aim to mimic complicated and visually detailed phenomena in nature.
• be able to use general and special purpose programming languages to generate images and certain types of geometry
• be able to describe and implement a selection of classic methods for procedural generation of patterns.
• carry out a project that solves a real problem that is formulated in terms of the desired visual result, where the student must be able to choose, motivate, implement, describe and to some extent evaluate their own solution based on existing, published methods.

Content

The principle behind procedural images, their use, advantages and disadvantages of texturing. Textures in 2, 3 and 4 dimensions. Regular and irregular patterns. Random number generation. Noise functions, especially "Perlin noise" and its relatives (lattice noises). Multifractal synthesis and simulation of turbulence. Fractals. Splines. Fractal Brownian Motion. Direct Fourier synthesis. Programming in general programming languages for procedural images. Shading languages. Functions for texture synthesis and geometry synthesis in modern graphics hardware. Introduction to deep learning methods.