Robust Detection of Significant
Points in Multiframe Images
-
multiframe SP detection, i.e. detection
in two or more images of the same scene which are supposed to be blurred,
noisy, rotated and shifted with respect to each other
-
an important pre-processing step in
image registration, data fusion, object recognition and in many other tasks
-
the point sets extracted from different
frames have relatively high number of common elements. This property is
highly desirable for further multiframe
processing.
-
the method exploits the intuitive understanding
of the "to-be-corner" property
A parameter approach was used to handle
differently distorted images. Points, which belong to two edges with
an appropriate angle in between regardless of its orientation are
understood here as significant points. Information about the number of
edges passing through each pixel and about the angle between them
is acquired from the number and distribution of local sign changes in the
difference between the image function and its local mean values.
Detection of significant
points in two different frames of the same scene: in the original (left)
and in the image blurred by 9x9 averaging mask and rotated by pi/9. The
significant points were detected by the Kitchen and Rosenfeld's method
(top), the Harris' method (middle) and by our method (bottom).
Other examples of detected
features:
-
registration of blurred images
-
registration of brain images
Relevant publications: