Abstract: This paper presents a non-parametric region competition scheme which combines scale-space clustering and region competition to segment the image. It also proposes a formal and general procedure to automatically find the initial regions. Our algorithm can segment an image into regions which are not homogeneous in the sense of statistics, but homogeneous in the sense of semantics with respect to the segmentation context. We call it semantically homogeneous segmentation of the image. Using both semantic homogeneity and quantitative control of the number of the resultant homogeneous regions, our algorithm may produce a 'clean' resultant image, thus simplifying the following procedures.