Image Processing Reference
In-Depth Information
In Figure 3.7, we consider an image with almost a bell-shaped graylevel histogram. Sep-
aration of dark and bright regions in this image is a non-trivial task as the histogram is
not well-defined for thresholding using many existing algorithms. It is evident from the
figure that the proposed bilevel thresholding methodology (algorithm (i)), algorithm (iii)
and algorithm (vi) perform much better than the others in separating the dark areas in the
image from the bright ones. An image of a galaxy is considered in Figure 3.8. The graylevel
histogram of this image is almost unimodal in nature and hence extracting multiple regions
from it is a non-trivial task. We use the proposed multilevel thresholding scheme and the
various other schemes to find out the total extent and the core region of the galaxy. It is
evident from the figure that the results obtained using the proposed thresholding method-
ology is as good as some of the others. While the 'white' shaded area in the result obtained
(a) The Image
(b) Graylevel Histogram
(c) Segmentation by (i)
(d) Segmentation by (ii)
(e) Segmentation by (iii)
(f) Segmentation by (iv)
(g) Segmentation by (v)
(h) Segmentation by (vi)
(i) Segmentation by (vii)
FIGURE 3.9: Performance of the various thresholding algorithms applied to segment a 'low
contrast' image into three regions
using algorithm (i) represent a region slightly larger than the core region, the 'white' shaded
Search WWH ::




Custom Search