Title: Tree: a knowledge-based system for generating images of trees
Abstract: This thesis deals with the design and implementation of a knowledge based system for generating a variety of tree images. Since it is meant to be used in a highly interactive design setting, the system must produce convincing images at high speed. It also has to allow the user effective and convenient parametric manipulation of generated trees.
To speed up the process of tree generation, the Direct Image Synthesis approach is introduced. This approach saves considerable amounts of time in bypassing the modeling phase of commonly used object based approaches. As it operates upon images and not 3D objects, the Direct Image Synthesis approach encourages the application of rule based principles of depiction analogous to those used by a painter. These depiction techniques prove to be an efficient alternative to computationally expensive rendering techniques in computer graphics.
The model uses Structured Recursion for construction of the tree image to make the modeling process more intuitive. Structured Recursion enables functional differentiation between the Trunk, Bough, Branch and the $Twig$ which are considered the principal elements in perception of a tree. This differentiation makes possible the implementation of Top-Down (perceptual) modeling process and strengthens the model's ability to depict a variety of species. The hierarchical structure of the model, which is conceptualized after the principles of abstraction in painting, follows directly from the application of Structured Recursion.
The parameters of the model encode those characteristics of a tree that define its form, texture and color. They are meant to be readily understood and to allow easy and versatile manipulation of the tree image. The thesis includes an in depth discussion on the definition and behavior of each parameter as well as on the relationship and the nature of dependency among them. The Chapter Setting the Parameters to Generate Types of Trees discusses the patterns of manipulating the tree image by gradual adjustment of parameter values and presents a series of tree images accompanied by the tables of their parameter values.
Preliminary research on methods to express distance in the composition of tree images has been undertaken. Leonardo da Vinci's rules from the Perspective of Disappearance and the Aerial Perspective are applied to the model. A comparative assessment of the Object Based and Image Processing Approach is also performed.
Publication Year: 1990
Publication Date: 1990-01-01
Language: en
Type: article
Access and Citation
Cited By Count: 4
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot