Color image segmentation based on intensity and hue clustering - a comparison of LS and LAD approaches

Abstract

Abstract. Motivated by the method for color image segmentation based on intensity and hue clustering proposed in [26] we give some theoretical explanations for this method that directly follows from the natural connection between the maximum likelihood approach and Least Squares or Least Absolute Deviations clustering optimality criteria. The method is tested and illustrated on a few typical situations, such as the presence of outliers among the data.

Cluster Detection in the Noisy Environment by Using the Modified EM Algorithm

“…Clustering is a widely used exploratory data analysis tool that has been successfully applied to data analysis, image processing, pattern recognition, engineering [2,4,6,7,8,15,17,18], and many other fields. In this paper, we focus on the detection of clusters in a noisy environment based on the well-known EM algorithm [2,3,9,11,18].…”

“…In this sense, the problem of dispersion is solved as the LS (Least Squares) and the LAD (Least Absolute Deviation) problems which have a wide variety of applications, such as image processing, clustering, data analysis, outliers detection, pattern recognition etc. [1,6,12,14,17,19]. To accomplish a better clustering quality, the dispersion of each cluster is pondered with a rejection parameter α > 0, where data is considered noisy if it exceeds each threshold.…”