Natural image segmentation is an important topic in digital image processing, and it could be solved by clustering methods. We present in this paper an SOM-based k-means method (SOM-K) and a further saliency map-enhanced SOM-K method (SOM-KS). In SOM-K, pixel features of intensity and L∗u∗v∗ color space are trained with SOM and followed by a k-means method to cluster the prototype vectors, which are filtered with hits map. A variant of the proposed method, SOM-KS, adds a modified saliency map to improve the segmentation performance. Both SOM-K and SOM-KS segment the image with the guidance of an entropy evaluation index. Compared to SOM-K, SOM-KS makes a more precise segmentation in most cases by segmenting an image into a smaller number of regions. At the same time, the salient object of an image stands out, while other minor parts are restrained. The computational load of the proposed methods of SOM-K and SOM-KS are compared to J-image-based segmentation (JSEG) and k-means. Segmentation evaluations of SOM-K and SOM-KS with the entropy index are compared with JSEG and k-means. It is observed that SOM-K and SOM-KS, being an unsupervised method, can achieve better segmentation results with less computational load and no human intervention.

Medical image segmentation plays an important role in medical visualization and diagnosis. We study in this paper an automatic segmentation method for liver magnetic resonance (MR) images based on the self-organizing map (SOM) and hierarchical agglomerative clustering method. At first, the local features of the MR image pixels are extracted to feed the SOM after a pre-processing step. The output prototypes are then filtered with the hits map and a hierarchical agglomerative clustering method is applied to the prototypes to select the best segmentation according to a quantitative image evaluation index. The segmentation results after the post-processing show the proposed method to be effective and promising. Further research work is also recommended.

This paper presents an image-based artistic rendering algorithm for the automatic Pointillism style. At first, ellipse dot locations are randomly generated based on a source image; then dot orientations are precalculated with help of a direction map; a saliency map of the source image decides long and short radius of the ellipse dot. At last, the rendering runs layer-by-layer from large size dots to small size dots so as to reserve the detailed parts of the image. Although only ellipse dot shape is adopted, the final Pointillism style performs well because of variable characteristics of the dot.

No abstract