Volume Rendering (3D VR)

Ray-casting a whole volume into a 3D image through a transfer function.

Volume rendering casts a ray through the reconstructed volume for every pixel on screen, accumulating colour and opacity from every voxel the ray passes. A transfer function (called a preset in dcmage) maps voxel intensity to that colour and opacity, and that mapping is what makes bone, lung or vessels the visible structure.

Unlike MPR, no pixel in the result corresponds to a slice: each one summarises a line through the data. It is a way to see the shape of anatomy, not a surface to read or measure on.

The abbreviation is unfortunate — in DICOM, VR normally means Value Representation, the two-letter code describing how a tag’s value is encoded. dcmage writes “3D VR” for the rendering and plain “VR” for the tag attribute.