Image Processing Reference
In-Depth Information
FIGURE 10.6: Segmentation comparison. (a)Test image140055(b)Histogram based seg-
mentation (c)Histon based segmentation (d)Roughness index based segmentation
A rough set theoretic thresholding algorithm has been presented here for segmentation of
color images. The histon, introduced in section 10.3, presents a new means for visualization
of color information in an image by exploiting the correlation among the pixels in all the
three color planes. The roughness index, obtained by correlating histon to the upper ap-
proximation and the histogram to the lower approximation of rough set, has been employed
for obtaining the optimum threshold values to yield better segmentation performance. The
visual and the quantitative comparisons demonstrate the eectiveness of the algorithm.
Search WWH ::

Custom Search