The display of multi-valued data in a single image is challenging. This new technique borrows concepts from painting to effectively display many values simultaneously. This image shows a 2D field of diffusion tensor data, a scalar intensity image, and the underlying sampling rate. The data have seven values at each location in the image, and the relationships among the values are important and difficult to understand otherwise. The image is built up in layers: intensity image (top left); sampling rate grid (top center); ellipse shapes showing in-plane component of principal diffusion and anisotropy (top right); texture on ellipses showing absolute diffusion rate (bottom). Right-left asymmetry in the diffusion tensor information shows regions of pathology in this diseased mouse spinal cord.

This image simultaneously displays all six values of this 2D magnetic resonance image of a mouse embryo. A viewer can see relationships that would otherwise be difficult to understand. Three of the values are mapped to red, green, and blue. A fourth value maps to opacity, showing the background color where the value is low. A fifth value maps to the size of small pointilist-esque ellipse shapes and a sixth to their direction. 

Slice of mouse spinal cord with MS-like disease. Left: 3x3 array of individual components of matrix representating tensor.  Right: tensor ellipsoids reflecting geometric structure within voxels. 

Volume rendering of 6-valued MR data of 12.5 day mouse embryo. Opacity and gray-level linear combinations satisfy mathematical goals: 1) create effective mask, 2) contrast materials well, and 3) keep noise at acceptable level. 

Semi-transparent volume-rendering of MRI data collected from a locust.