Direct cell-to-cell volume visualization algorithms, also called proje
ctive rendering methods, present the advantage of allowing semi-transp
arencies and, additionally, as they project all the samples, they avoi
d the voxel-space aliasing. However, they are generally computationall
y expensive and artifacts may appear in the projection. In this paper,
different projective strategies-are reviewed and compared. A new algo
rithm, based on a back-to-front (BTF) voxel-to-voxel traversal, is pro
posed. It consists of the computation, in a preprocess, of 16 Extended
Generalized Weight (EGW) matrices, based on the voxel's geometry and
the viewpoint direction. Taking into advantage the voxel-to-voxel cohe
rence, the EGW are applied incrementally to each voxel as footprints t
o be composited with the image buffer, according to the voxel shade an
d opacity values. It is shown that the method avoids artifacts, and th
at it has a low computational cost because it does not require re-voxe
llization and it does not perform any geometric computation in the pro
jection process. It is specially suitable for voxel representations an
d image space resolutions such that the projection of a single voxel c
overs more than one pixel.