Accelerated intuitionistic fuzzy clustering for image segmentation

To improve processing time of the intuitionistic fuzzy C-means during color image segmentation, this paper proposes a scheme based on two clustering stages. In the first, a downsampled image is used to isolate the dominant color of the images by means of centroids calculating. Later, in the second stage these centroids are used during the image segmentation. With these two processes, an algorithmic acceleration of approximately eleven times can be guaranteed compared to the conventional algorithm. The effectiveness of this proposal is verified by experiments on the natural color images of datasets such as BSDS500 Alpert et al. Segmentation Evaluation Database, Sky dataset, Stony Bro- ok University Shadow and ISIC 2018. The quality of the segmentation was quantified using metrics and compared with other current methods of the state of the art. The results obtained show a superior performance of the proposed method both in segmentation and in processing time.