Rough segmentation of natural color images using fuzzy-based hierarchical algorithm

This paper proposes rough segmentation of natural color images using fuzzy-based hierarchical algorithm. Statistical Geometrical Features (SGF) are adopted as texture features and L*a*b* color space is used to represent a color feature. Fuzzy homogeneity decision makes a fusion of texture features and color features. Proposed hierarchical segmentation method based on the fuzzy homogeneity decision is performed in four stages: hierarchical splitting, local agglomerative merging, global agglomerative merging and pixelwise classification. Experiments on segmentation of natural color images are presented to verify the effectiveness of the proposed method in obtaining rough segmentation.