Image Processing Reference
In-Depth Information
(12.5b)
Im15,.961
(12.5c)
(12.5d)
Im19,.928
(12.5e)
Im11,.923
(12.5f)
(12.5a) Im3,1.0
Im5,.946
Im6,.921
(12.5g)
(12.5h)
(12.5j)
(12.5k)
Im18,.856
(12.5l)
(12.5i) Im7,.896
Im8,.906
Im9,.903
Im10,.866
Im20,.839
(12.5m)
Im13,.834
(12.5n)
Im14,.826
(12.5o)
Im17,.826
(12.5p)
Im16,.819
(12.5q)
Im12,.773
(12.5r)
Im4,.677
FIGURE 12.5: Images ordered by tNm-measure values
12.6
Conclusion
This article focuses on an approach to solving the image correspondence problem by con-
sidering description based a
nities between coverings of digital images. It introduces an
approach to measuring the resemblance between pairs of images using several L
norm-based
tolerance nearness measures. The proposed solution to the image correspondence problem
is based on a comparison of the descriptions of elements of tolerance classes contained in
coverings of pairs of digital images. It should be noted here that the proposed approach
to measuring similarities between perceptual granules is not limited to digital images. The
proposed approach also has promising implications for segmenting videos, especially in ap-
plications where grouping images in a video depends on very refined similarity measurements
over many separate images contained in a video.
2
The author is deeply grateful to James F. Peters and Amir H. Meghdadi for suggestions and
insights concerning topics in this chapter. I especially want to thank Amir H. Meghdadi for
the use of his tolernance nearness measures toolset. This research has been supported by
the Natural Sciences and Engineering Research Council of Canada (NSERC) grant 194376.
Search WWH ::
Custom Search