International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
Mean shortening of edges at point 20
10 — ———— 1 T 1 T^ Ir À4—7*]
+ *
= ++
a 5r4 + 3 4
= + + + + +
+ 1 E FT 4
} 24
+" pt
ol EL. tT ld 1 + ui +++ts+ 1 1 =
0 5 10 15 20 25 30 35 40 45 50
Starl RE UNRation
8 T T T T T T T T
+
6r xl
T +
© p M x
2 t For ; p ; x =
Qd edepol E sb spese TE Sh ES
0 5 10 15 20 25 30 35 40 45 50
Image No.
Figure 8: Shortening of straight lines and edges at a single junc-
tion over all images
Mean shortening at each point over all images
157 T T T T
|
|
ak 10+ fy
& | + + + ++
: | + 4 Ë
= 5r rt E p Rr tr, * ert s «zu m meto ++ =
| +
| + + + +
0 E 1 1 i L
0 10 20 30 40 50 60
Stanialü diQiation
6 f T T T
| i a + + 4
at t * To D TE |
= | + Eun + E + +
= | t BL rs *
o i + +
2} d
|
ol À i L L 1
0 10 20 30 40 50 60
Point No.
Figure 9: Shortening of straight lines and edges for all junctions.
Mean and the variance are taken over all images
role of reference data in characterizing and evaluating algorithms
is discussed and the approaches to reference data generation are
explained.
The methods were successfully employed to investigate the noise
sensitivity of point extraction modules and the shortening of straight
lines and edges provided. by linear feature extraction modules.
First results are plausible and give reason 1) to exploit the pro-
posed methods for reference data generation on other data sets,
2) to further investigate the noise behavior of corner extraction
modules including local image characteristics such as shape or
local contrast and 3) to further investigate the shortening of linear
features dependent at junctions dependent on the contrast at the
junction branches.
ACKNOWLEDGMENT
This work has been supported by the German Research Council
(DFO).
References
Canny, J.F., 1983. Finding Edges and Lines in Images. Technical
report, MIT Artificial Intelligence Laboratory.
Crowley, J.L., Riff, O. and Piator, J.H., 2002. Fast computation of
characteristic scale using a half octave pyramid. In: CogVis
2002, International Workshop on Cognitive Vision, Zürich.
Fórstner, W., 1996. 10 Pros and Cons Against Performance Char-
acterization of Vision Algorithms. In: Workshop on "Perfor-
mance Characteristics of Vision Algorithms" , Cambridge.
1066
Fórstner, W. and Gülch, E., 1987. A Fast Operator for Detection
and Precise Location of Distinct Points, Corners and Cen-
tres of Circular Features. In: Proceedings of the Intercom-
mission Conference on Fast Processing of Photogrammetric
Data, Interlaken, pp. 281—305.
Fuchs, C., 1998. Extraktion polymorpher Bildstrukturen und ihre
topologische und geometrische Gruppierung. DGK, Bayer.
Akademie der Wissenschaften, Reihe C, Heft 502.
Luxen, M., 2003. Variance component estimation in performance
characteristics applied to feature extraction procedures. In:
B. Michaelis and G. Krell (eds), Pattern Recognition, 25th
DAGM Symposium, Magdeburg, Germany, September 10-
12, 2003, Proceedings, Lecture Notes in Computer Science,
Vol. 2781, Springer, pp. 498—506.
Maimone, M. and Shafer, S., 1996. A Taxonomy for Stereo Com-
puter Vision Experiments. In: ECCV Workshop on Perfor-
mance Characteristics of Vision Algorithms, pp. 59 — 79.
CA
KEY
ABS
With
auto
ary c
COTTE
came
angl
can |
estal
quan
subs
the c
Mot
man
came
proc
tions
racy.
for s
the r
proa
with
solut
and
deter
App
eran
sic c
parai
terio
the c
to th
resp
calle
2000
the c
and :
tion :
came
ploit
indus
enab
vatia
mate
appri
