Illuminated 3D Scatterplots
(c) 2009 H. Sanftmann, D. Weiskopf


This videos demonstrate the usefulness of our technique. They show a direct comparison of our approach (on the right side) with traditional 3D scatterplots.




UFAC This video is showing the covtype dataset. Given the illumination samples belonging to different regions can be distinguished and structures can be recognized.







UFAC This video is showing the bucky-ball dataset. The first and second derivative and the scalar value are mapped to the three axes of the plot. With our 3D scatterplots it is clearly visible that the values are distributed in two separate planar regions. We also show the effect of manipulating the weights resulting in a modified classification of samples.