Professional Documents
Culture Documents
Corner Detection Using Color Density Gradient: Sun Da Huang Jianhua Tang Xianglong
Corner Detection Using Color Density Gradient: Sun Da Huang Jianhua Tang Xianglong
Corner Detection Using Color Density Gradient: Sun Da Huang Jianhua Tang Xianglong
Gradient
Sun Da1 Huang Jianhua1 Tang XiangLong1
1
School of Computer Science and Technology,
Harbin Institute of Technology, Harbin China
(a) (b)
Fig.4. (a) The corner points located in
density gradient based detector. (b) Cor-
ners located by intensity gradient based
detector. For two detectors, ‘+’ is with the
detection scale σ = 1, and ‘o’ is with the
detection scale σ = 4.
Fig.2 The corners detected by the DGBD
in color image. In new corner detection method, the
density gradient is used to take place of
the intensity gradient while compute the
Harris corner response. Because Harris
corners are invariant to the rotation
change, the scale stability of density
gradient based method is compared
with that of the intensity gradient based
method. In [18] Schmid uses the re-
peatability as the criterion to evaluate
Fig.3 The corners with the color of the the interest point detectors and it is
building detected by DGBD. widely adapted to evaluate the new de-
tectors. The repeatability score defined
3. Evaluation by Schmid [18] is the ratio between the
number of correspondence points and
In large detection scale, the noised and the smaller number of points detected in
textures will be filter out, and the impor- one of the images. The set of corre-
tant objected information are enhanced. spondence points Ri (ε ) and the new re-
Therefore, we are more concerned about
the corners detection in large scale. The peatability ri (ε ) is defined as (6) and (7):
density gradient based detector can pre- Ri (ε ) = {( x1 , xi ) | dist ( H1i x1 , xi ) < ε } (6)
serve the corner position while density
Ri (ε )
estimation scale increasing, as Fig.2 (a) ri (ε ) = (7)
shows, because the density estimation is min(n1 , ni )
realized by smoothing in the joint spatial- where, H1i is the homograph between im-
feature space, and Darin [4] illuminated ages I1 and Ii. ε is the max distance of two
4. Conclusions
60
40
20
0
1 1.5 2 2.5 3 3.5 4 4.5
image scale change
5. References