Digital Image processing deals with manipulation of digital images through a digital computer in order to get an enhanced image or to extract some useful information from it. GIMP and Adobe Photoshop are the most common examples. They are the widely used application for processing digital images. There are many limitations of traditional methods of processing digital. Hence there is a motivation to use fuzzy logic for digital image processing. This article throws light on reasons for using fuzzy image processing methods compared to traditional digital image processing methods.
1. Problem with Traditional Image Processing
In digital image processing, it may not be appropriate and accurate to define quantities such as dark or light using crisp set theory. For example, if someone wants to define a set of gray levels that has the property dark. In crisp set theory, a threshold is determined; say it is gray level 100.
Figure-1 shows that, all pixels whose gray levels are between 0 and 100 belong to set dark and other pixels do not. pixel value 99 belong to dark set only. But the darkness is a relative measure. Hence, pixel value 99 may also belongs to other set. Thus, traditional image processing methods fails to handle such situations.
2. What and Why Fuzzy Image Processing
Fuzzy image processing, which is based on fuzzy set and/or fuzzy logic, is the collection of all approaches that understand, represent and process the images, their segments and features as fuzzy sets. Many difficulties in image processing arise due to ambiguity and vagueness in data/tasks. There are imperfection in the image processing such as greyness ambiguity, geometrical fuzziness, and ill-defined knowledge.
The real-time image processing problems are fuzzy in the nature. The questions such as (a) whether a pixel should become darker or brighter than it already is?, (b) where is the boundary between two image segments?, (c) what is a tree in a scene analysis problem?, and other similar questions are examples for circumstances that a fuzzy approach can be the more suitable way to manage the imperfection.
Fuzzy set and Fuzzy Logic is one of the sub-area of Artificial Intelligence to deal with such imperfections in digital image intelligently.
3. How Fuzzy Solves The Above Problem
A fuzzy set can represent can represent the dark set problem easily. As shown in figure-1 (b), two thresholds may be defined; gray levels 50 and 150.
All gray levels that are less than 50 are the full member of the dark set. All gray levels that are greater than 150 are not the member of the set. The pixels, whose gray levels are between 50 and 150, have a partial membership in the set.
Thus pixel value 99 partially belongs to dark set and partially do not belong to it. This is how fuzzy based methods overcomes the problem of traditional image processing.
Therefore, if digital images have are imperfection such as grayness ambiguity, geometrical fuzziness, and ill-defined knowledge, Fuzzy Based Methods for Image Processing
- 1. H. R. Tizhoosh. Fuzzy Image Processing. Springer-Verlag, October 1997.
- 2. R. Krisnapuram J. C. Bezdek, J. Keller and N. R. Pal. Fuzzy Models and Algorithms for Pattern Recognition and Image Processing. Springer, 2005.