FCM Algorithm: Analysis of the Membership Function Influence and Its consequences for fuzzy clustering

Mantilla L.; Yari Y. · 2020 IEEE Colombian Conference on Applications of Computational Intelligence, ColCACI 2020 - Proceedings (2020) · DOI: 10.1109/colcaci50549.2020.9247944

Resumen

Image segmentation in satellite images is a task widely investigated since we can extract some information of an image and analyze it. We propose to use a weighted factor for each of the distances used to calculate the degree of membership of each element to the cluster. In this way, we seek to reduce the influence of the upper and the lower bounds on the FCM equa. tion. This paper reports preliminary results of the experiments and shows that the proposed algorithm performs accurately on a real dataset. For the evaluation of the algorithm, different cluster validity indexes are employed.

Instituciones: Universidad Católica San Pablo; Universidad Católica de Trujillo Benedicto XVI