Edge-Enhancing of Color Segmented Images

This paper introduces a new approach to edge-preserving smoothing of color segmented images. It is well suited for an efficient reduction of the texture contours in colored images that is an essential step in applications such as image classification or image retrieval. The algorithm reduces the artifacts in the homogeneous areas, but preserves all image structures like edges or corners. Our procedure uses mean-shift algorithm to obtain a colored segmented images. The filtering algorithm is not only applicable to color images segmented using mean-shift, but can be applied to any segmented images too. It is formalized through the definition of a morphological window and a homogeneity measure on this window. The adaptive filter presented here tries to overcome some of the disadvantages of existing smoothing filters and is conceived as an extension of the morphological profiles filtering. The experimental results over Berkeley database images show that the proposed method is well suited for textured image scenes.