Botanical Visualization of Huge Hierarchies

The program has been developed after the method of Ernst Kleiberg, Huub van der Wetering und Jarke J. van Wijk presented in their paper "Botanical Visualization of Huge Hierarchies". It describes how to visualise the arrangement of data structures by mapping nodes to branches, their children to subbranches and sets of leaves to fruits.

Implementation

The application has been written for Studierstube. It mainly consists of two sets of classes: SoTreeBase and inheritances to describe the input hierarchical structure and SoBotanicalBase and inheritances to describe the structure of the botanical tree. The SoBuildBotanicAction maps the input hierachical structure into the botanical tree, and sensors coupled to the fields of each botanical branch update the fields of their subbranches to determine all the geometric transformations (rotations, translations and scalings) of the elements of the botanical tree.

Installation

The following requirements are necessery for the installation:

• Studierstube Version 3.2
• CVS checkup stbapi, PucPipLayout

How to use the program?

To use the application you first need to specify the information of the hierachical structure in the input file. Here are two examples of such input files:

After loading this input file a GUI gets presented to you. You can change the length of the branches, a scaling factor for the thickness, the rotation angles alpha and beta, and the number of branch connections. To rebuild the tree you always need to click on the push button "build tree".

Images

Screenshots generated with BotanicalTree1.iv:

branch connections = 2; alpha = 0.9	branch connections = 2; alpha = 0.4
branch connections = 3; alpha = 0.9
Screenshots generated with BotanicalTree2.iv
branch connections = 2; alpha = 0.9	branch connections = 2; alpha = 0.4
branch connections = 3; alpha = 0.6

Improvements

There are some aspects from the paper that haven't been implemented. Here are some ideas to improve the application:

• The fruits can visualise more information when taking the geometry of Phi-Balls where each leaf is represented as a cone on the surface of the sphere.
• The geometry of the extrusions can be improved by defining these with spline functions.
• To make the visualisation work with real data, a parser can be implemented that scans the local disk and automatically generates the input file for the application.
• By means of more interactivity, information concerning the actual position in the hierarchy or the contents of the leaves can be shown.