FAST TEXTURE-BASED TENSOR FIELD VISUALIZATION FOR DT-MRI

Diffusion tensor MRI (DT-MRI) provides useful connectivity information about neuronal tissue, but has several visualization challenges. The high dimensional nature of the data (a 3D volume image with a matrix at each voxel) requires some processing to visualize on conventional 2D display devices. In this paper we present a novel approach to visualizing the tensor-valued data. Our approach requires minimal preprocessing and can be implemented and run much faster than other existing texture-based approaches. This is accomplished by using a hybrid approach, which has similarities to existing glyph, particle and texture-based visualization techniques. The resulting image is formed by blending textured billboards which are rendered at the center of each voxel in the dataset. Color, texture orientation, texture frequency and animation are used to convey diffusion properties. Results are presented for a DT-MRI image of a human brain.