Application of fuzzy logic techniques to image segmentation

Segmentation is one of the most important preprocessing operations that are used in image system analysis for image understanding and image Vecognition.Since most images have many different regions, image Segmentation algorithms* have been developed for extracting these regions, 'image segmentation process is divided into two main classes; Discontinuity Detection and Similarity Detection. The objective of this thesis is the application of Fuzzy Logic techniques to constructing general-purpose Image Segmentation Algorithm for both classes. Gradient Operator technique has been implemented for comparison with Fuzzy Logic technique in Boundary Detection. Two approaches of Fuzzy Logic have been developed. The first one uses Single Fuzzy Set. The Set in this case was the Difference. The second approach uses Double Fuzzy Set, which are the Difference and the Standard Deviation. Simulation result showed that the Fuzzy methods was superior to the Gradient method and Double Fuzzy Set approach gave good result independent on the image used. In the second class (Similarity Detection), a well known method for Split and Merge has been implemented for comparison purposes. This method uses single merge operation. To reduce the number of regions contained in the segmented region, a multiple Fuzzy Merge method has been developed. Results obtained by this method
were encouraging.