Kernel Regression from scratch in python
Implementing kernalized regression from scratch in python was more challening than anticipated. So i wrote a summary that explains the code. I also tried a different writing style and originally published this article on medium.com.
Synthetic Silviculture: Multi-scale Modeling of Plant Ecosystems
Creating geometry for large scale Ecosystems in Houdini to create photorealistic ecosystem images
Recreating realistic bokeh by biasing lens-sample positions
Ex 6.4: Unlike real lenses most ray-tracer only consider a perfectly disk-shaped circle of confusion. However the distorted circle of confusion known as bokeh is a key element in many photographs. As such i present a method which approximates this distortion in the pbtr-system
Partical fluidsimulation based on double density relaxation
The simulation is based on a double density relaxation scheme to model the pressure gradient. As it borrows ideas from position based dynamics solid and viscoelastic objects and the interaction between them could easily be integrated into the simulation.
Fast ray-AABB Intersection test
EX 4.9: Here i implemented the fast-AABB test proposed by Eisemann 07 and analyse it's impact on the performance of the pbrt-system.
A simple Game
In the game the player controls a marble which he has to navigate through a obstacle course. The Game was written entirely in C++ using openGL features like textures, normal Maps and HDR-tone-mapping. A custom export script allows the scene-segments to be modeled in the blender-editor and then to be loaded into the game. Running on both Windows and Linux
L-Systems a simple Tree
Data Oriented Design
Short Essay on Data Oriented Design focusing on CPU-Cache utilization. Furthermore some downsides of the most eminent and thought programming paradigm Object-Oriented-Programming are investigated