A study on performance improvement of Greg Turk's Algorithm on texture synthesis on surfaces

• Main Authors: Bongato, Paulo Napoleon S., Paguyo, Mark Simpson A., Ponce, Jose Mari N., Solis, Earl S.
• Format: text
• Language: English
• Published: Animo Repository 2008
• Subjects: Computer Sciences
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/14412
• Institution: De La Salle University

Procedural generation, also known as procedural synthesis, is using processing power and algorithms in order to automatically produce contents which include textures and animation. Procedural generation is an approach that could be used in content-dependent software such as games and image tools. It uses symmetrical properties of objects which could be described mathematically to reproduce repeating patterns using random-seed generation. This could reduce labor costs among artists and developers. There are different algorithms available that perform procedural texture synthesis. These algorithms are applicable only to specific applications. However, these algorithms have some issues to resolve or paths to improve which include the language it uses and the speed and evaluation of creating textures. The proponents of this research would like to investigate on the possible improvements of the performance of an algorithm presented by Grek Truk on aTexture Synthesis of Surfaces. This algorithm uses mesh hierarchy, vector fields to create the textures from a sample image and placing the texture at the same time on a specific surface. Part of the research would be the identification of the bottlenecks with the algorithm and identification of possible solutions using alternative data structures and algorithm.