A Comparison of Fuzzy C-Means and K-Means Clustering for Extraction of City Colours

Abstract

Abstract: The colour palette of each urban metropolis reflects its cultural identity and its unique flair. For urban development and urban renewal projects, incorporating a city’s existing colour palette into a construction plan would ensure that the completed project would be in harmony with the existing colour schemes of a neighbourhood. An earlier investigation implemented Fuzzy C-Means (FCM) colour extraction to identify five dominant colours from the images of each of the twelve major cosmopolitan cities that are situated on six continents. These cities were Cairo, Cape Town, Singapore, Tokyo, Perth, Sydney, London, Madrid, Mexico City, New York City, Buenos Aires, and Lima. This current research is a follow-on investigation of the previous city colour extraction study. It applies K-Means clustering to acquire five dominant colours for each of these 12 cities. The goal of this current research is to determine if the type of clustering method implemented would impact the results of the extracted colours. This is accomplished by comparing the city colours obtained by executing the K-Means clustering technique on these 12 cities and comparing these colours with those that were obtained from the previous FCM study