Speed Parameters in the Level-Set Segmentation

In image segmentation, based on the level set method, the evolution of the curve is determined by the speed function. In this paper we apply the level set segmentation to speckled images, in particular SAR (Synthetic Aperture Radar) images. Moreover we propose a parameters tuning of the speed function, obtained from the linear combination of the speed function of average intensities and of the image gradient. To show the validity of the proposed approach, we compare the segmentation results obtained from both synthetic and real images. Since there are not benchmark SAR images, computer images are been synthesized using speckle noise. Thus we show that the proposed speed function produces the best results, tuning the parameters in opportune way. The SAR images are PRecision Images (PRI), acquired during European Remote Sensing (ERS2) mission and CosmoSkyMed (CSM) image.