Fuzzy Sets
What is FIP
Why FIP?
History
Theory
Examples
Literature
Software
Conferences
Related Links
Contact

Fuzzy Edge Detection

There are different possibilities for development of fuzzy edge detectors:

Definition of appropriate membership funcions

This is the easiest way. One can define a membership function indicating the degree of edginess in each neighbourhood (Tizhoosh, 1997):

Edge detection using a simple membership function (Adapted from: Tizhoosh, Fuzzy Image Processing, © CopyRight Springer,1997)

This approach can only be regarded as a true fuzzy approach if fuzzy concepts are additionally used to modify the membership values. The membership function is determined heuristically. It is fast but the performnace is limited.

Rule-based fuzzy edge detection

Using appropriate fuzzy if-then rules, one can develop general or specific edge detectors in pre-defined neighborhoods:

Edge detection using a rule-based approach (Adapted from: Tizhoosh, Fuzzy Image Processing, © CopyRight Springer,1997)

See: Russo, F., Ramponi, G. (1994): Edge extraction by FIRE operators. In: Proc. 3rd IEEE International Conference on Fuzzy Systems, Band I, S. 249-253.

Other fuzzy approaches

One can also use other fuzzy concepts and classification techniques.
Example: FEDGE - Fuzzy Edge Detection by Fuzzy Categorization and Classification of Edges ( by Kenneth H. L. Ho, Noboru Ohnishi, Japan)

Content by:
H. R. Tizhoosh

Created by:
Log Web Design

Powered by:
PAMI Lab

Created: June 1997
Updated: Nov 2004