Comparative study of performance of distance measures in fuzzy C means clustering for CIELUV color images
Packyanathan Ganesan, V. Rajini, Vladimir Kalist, P. Vamsi Krishna · 2014
Segmentation is one of the initial but vital processes in most of the image analysis and interpretation applications. In the image segmentation, the test image is divided into number of sub images called segments or clusters. All the pixels in the same segment having the similar characteristics such as texture, color or intensity. In this paper, the performance of squared Euclidean, city block and cosine distance measures in fuzzy c-means clustering is compared for the segmentation of satellite images. This experiment is performed in CIELUV color space which has many advantages as compared to RGB color space. The proposed method is tested with number of satellite images.