Image Processing Reference
In-Depth Information
With the emergence of fuzzy set and its significance, it is widely used
in diverse applications such as medical image processing, remote-sensing
application, pattern recognition and decision-making problems. The raw
medical images are blurred and unsharp with vague structures and unclear
boundaries. Bezdek [2] used fuzzy models for segmentation and edge detec-
tion in medical image data. Images are in a sense 'fuzzy', and the task is to
reduce the fuzziness present in the image to obtain a clear image with sharp
boundaries. So, the task of medical image processing is not only to fuzzify
the image (i.e. to consider the vagueness present) but also to defuzzify the
image (i.e. to obtain a clear image).
While attempting to process medical images for analysis, it may be a good
idea to consider the fact that a computer vision system is usually embed-
ded with uncertainty and vagueness which needs to be appropriately taken
care of. Let us now see the logical reasons behind the applicability of fuzzy
notions in image processing by Prewitt [19]:
1. Images are 2D projections of the objects in a 3D world, and so some
information is lost while mapping.
2. In grey-level images, uncertainty exists due to the variability of grey
values. The pixel values in digital images are considered imprecise
and should be viewed as fuzzy numbers.
3. Many of the image definitions, such as boundaries/regions of
objects or the homogeneity of the segments or the contrast between
the objects and the background, require fuzzy notions for their
characterization.
4. Ambiguity is often present in the final interpretation of the image. Since
human understanding is never crisp or precise, it is important to incor-
porate soft decision-making since a hard decision may often be costly.
Fuzzy image processing is a collection of different fuzzy approaches to
image processing that can understand, represent and process the images.
It has three main stages - fuzzification (membership plane), modification of
membership values and, finally, defuzzification - as shown in Figure 2.1.
Input image
Output image
Image
defuzzification
Image
fuzzification
Membership
modification
Fuzzy theory
FIGURE 2.1
The general structure of fuzzy image processing.
Search WWH ::
Custom Search