19
the volume appears on the screen, regardless of input size, the faster the rendering
will be, which is as expected. Our results are taken from the slowest rendering time
for each of the test sets. Our method maintains a reasonable frame-rate even under
classification.
Figure 2.5 : A close up example of rendering using our piecewise tri-linear represen-
tation.
(a)
(b)
Figure 2.6 : A close up example of rendering using our piecewise tri-linear represen-
tation.