Digital Signal Processing Reference
In-Depth Information
→
Figure 4.24.
Q
results
of
section
6.
for
coastguard
(frames
30-39)
and-container
→
sequence
(frames 0-45,
downsampled by 5).
Each
point
represents
the
Q
for
each
video object plane to the ground truth video object plane. The circles show the
membership of the
→
Q
to the specific
video
object.
Interpretation
of
results
are
in
Section 4.23.
segmentation of the smaller boat. However, the white flagpole occludes
the container ship and split it into two segmented objects: one of poor
quality, and the other which can be considered a good quality partial
result. Overall, the good clustering of points shows the good controlla-
bility of the surface optimization algorithm w.r.t. the time smoothing.
Unfortunately, there does not exist ground truth data for the hall mon-
itor sequence so we cannot do this type of analysis.
Search WWH ::
Custom Search