Authors
Doria Jianu
Wenjin Zhou
Çagatay Demiralp
David H. Laidlaw
Abstract
We present a new method for visually differentiating integral curves obtained from DTI volumes. The goal of our method is to reflect the spatial relations between visualized integral curves. This will enable the user to group similar curves together while differentiating dissimilar curves that may lie near one another. To this end, we first define a similarity measure and compute the similarities between the curves according to that measure. Then we encode this similarity information onto the streamtube representations of the curves as stripe texture patterns. We present examples of visualizations using four different pattern styles. We evaluate our method with help of neuropsychologists; initial feedback suggests that our method can be useful in understanding the underlying connectivity of integral curves.