A level-set model driven by Bidimensional EMD for sonar image segmentation

This paper proposes a new multiphase level set model. Its energy function is driven by Bidimensional EMD (Empirical Mode Decomposition) to resolve the segmentation problem of sonar image. We introduce the EMD and BEMD, and give the steps of BEMD. It is used to extract intrinsic components of images. Then, we integrate them into the VC´s (Vese-Chan) multiphase level set energy functions to resolve the sensitiveness of level set models to noise. Experimental results show that the segmentation results of our method is superior than the VC level set model.