Data driven bandwidth for medoid shift algorithm
ICCSA'11 Proceedings of the 2011 international conference on Computational science and its applications - Volume Part II
| Hi-index | 0.00 |
In this paper, we present a new approach for automatic color image segmentation. It is a multilevel clustering method based on a new proposed non-parametric clustering algorithm, called adaptive medoidshift (AMS) and normalized cuts (N-cut). The AMS algorithm is a modification of recently presented medoidshift algorithm by transforming its global fixed bandwidth to local automatically chosen bandwidth for every data point. The AMS method locally clusters the image color composition by considering their spatial distribution, resulting into uniform segments. Then the segmented regions are represented by graph structure and finally N-cut method performs optimized global grouping into meaningful salient regions that convey semantic information of image. The experiments show that proposed segmentation method provides good segmentation results on variety of color images.