Image Processing Reference
In-Depth Information
g = 15 : n
f
= 118
g = 10 : n
f
= 243
g = 6 : n
f
= 588
g = 3 : n
f
= 2367
FIGURE 7.6: Isothetic covers of
Stanford Bunny
for different grid sizes;
n
f
= number of cover faces defined as isothetic polygons. (See color insert.)
Reprinted from
Proc. 14th International Workshop on Combinatorial Image Analysis: IWCIA11
, LNCS
6636
: 70-83,
N. Karmakar et al., Copyright 2011, with permission from Springer.
7.1.2 Test Results
A set of results by the algorithm on the object
Stanford Bunny
1
is pre-
sented in Fig. 7.6. It is quite apparent that as the grid size g is decreased,
a tighter approximation of the object is obtained; consequently, the number
of vertices, edges, and faces increase in a quadratic manner. In 3-D, at most
eight UGCs can be incident at a grid vertex. Depending on the object occu-
pancy of these UGCs, it is decided whether the grid vertex is a vertex of the
isothetic cover. For instance, if all eight UGCs have object occupancy, then
the grid vertex lies within the object, and hence it is not a vertex of the cover.
Based on the object occupancy and the nature of arrangement of the occu-
pied UGCs, the vertices can be classified into five categories: Type 1, Type 3,
Type 4, Type 5, and Type 7 [1], the denotation corresponding to the number
1
Source:
http://graphics.stanford.edu/data/3Dscanrep.