from wikipedia
Final project proposal
蘇彥禎
Problem
I plan to implement a volume rendering system in the final project of CG course. The system will map 3D scalar data to color and opacity value and integrate them to the final result as figure below. A big problem of volume rendering is how to find a suitable mapping method to create meaningful classification. I want to try whether the concept of multidimensional indexing mentioned in MMAI course can provide a new approach.
from wikipedia
This is my
current result.
This is the
lit result
Idea
1. Treat every voxel of the input volume as a 4 dimensional data that includes 3D position and intensity value.
2. Use MDS (Multidimensional Scaling) method to display the relationship of
these voxels.
I hope the distribution map can be as good as the example
below.
3. Try to design some operation for user to set the transfer function that maps
scalar data to RGBA value.
LLE example
Related work
The most popular classification approach is Joe Kniss's navigation method.
"Interactive Volume Rendering Using Multi-Dimensional Transfer Functions and
Direct Manipulation Widgets"
Current results
clustering interface
rendering result
