Not like scanning statues, we only have one chance to obtain all the
scan data, since the model is live and cannot return to the original
pose exactly after several days. It makes our experiment design
critical. The key problem is how to help the model keep stable
during the long scan, and return to the original pose for the next
scans after rest.
Firstly, we need a professional model who can stand more stable
for a long time. We ask many friends around us, contact the people
of art school, post an advertisement on the website Craiglist, and
finally choose one from tens of responses. We choose to do experiment
at night in the Room 6LW3, where has enough space and
allows us to avoid possible disturbing.
Two tripods are adjusted to proper distance and heights to rest the
arms and reduce the occlusion. The position of feet is marked at
the beginning of the scan. Four scan views are planned around the
model from frontal-left, frontal-right, back-left to back-right. The
distance from the scanner to the model ranges from 2.5 to 3 meters,
and the scanner is set as high as the shoulders. Ten targets are
placed around the subjects and properly named for the registration
usage.
In addition, a D-SLR camera Canon 20D is set up to monitor the
model movement. We keep comparing the images during one scan
or between different scans to make sure there is no obvious movement.
Scan precision is set to 2mm, so that it shall take less than 15 minutes
to complete one scan. Though we only have four scans, the
whole scan lasts for four hours because it takes a long time to move
the Leica Scanner and set it up in a new position. That’s what we
We show the four scan data in Figure 2. Vertices numbers of these
initial scans are huge, ranging from 90K to 150K. The high resolution
data pose a challenge to our following data processing.