5.4 Classification

The Levoy's equations described in Section 3.4, were used for classification to deduce a scalar field of opacities, $O[x,y,z]$. In addition, the application provides classification for two density values, $\rho_{v1}$ and $\rho_{v2}$, so that up to two surfaces can be seen in the final image. For example, this allows extraction of the the skin and the skull of the visible human head CT-scan for display in a single image. The density corresponding to the skin is given a semi-transparent opacity, $\alpha_{v1} = 0.25$, where as the skull is given an full opacity, $\alpha_{v2} = 1$. This process makes it possible to see the skull under a semi-transparent skin. A classification array of voxels, $A[x,y,z]$, indicates to which category each voxel belongs, i.e. skin, skull or nothing. Therefore, the computed values of different density can be assigned, either to the red, green or blue components of the image pixels. In this manner, the skin and the skull are rendered with different colour components allowing the viewer to better differentiate between the two surfaces. It could be easily extended to many more surfaces by keeping a table of opacity for mapping density value.