The goal of the suggested research was to see how well fuzzy C- means and competitive agglomeration combined for image segmentation. It's a vector quantization approach based on fuzzy clustering. Image segmentation is the process of dividing a digital image into several segments. Various strategies based on conventional approaches have been developed for segmentation, including region rising, threshold procedure, and watershed transform. Segmentation based on clustering techniques has emerged as a result of the disadvantages of these methodologies. Data clustering is based on the premise that each cluster is represented by its centroid. In addition, each cluster is represented by the similarity of the input vectors to the centroid. There are two types of clustering methods: parametric and nonparametric. Using a Euclidean distance between samples to find natural groupings in a dataset is a non-parametric technique. Non-parametric clustering techniques include K-means, hierarchical, and spectral clustering. One of these approaches' flaws is their inability to withstand image noise. As a result, a fuzzy segmentation technique is extensively utilised in image clustering and segmentation. Determining the number of clusters and selecting an objective function are the most difficult aspects of fuzzy c-means. As a result, we employ a vector quantization approach based on fuzzy clustering. A specialised objective function that combines fuzzy c-means and a competitive agglomeration term is used in this approach. The rebuilt images are of good quality, and the process is quick.
Author(s) DetailsI. Nandhin
M.E Communication Systems, Saranathan College of Engineering, Tiruchirapalli, Tamil Nadu, India.
V. Mohan
Saranathan College of Engineering, Tiruchirapalli, Tamil Nadu, India.
View Book :- https://stm.bookpi.org/AAER-V7/article/view/961
No comments:
Post a Comment