Multiphase segmentation based on new signed pressure force functions and one level set function

In this paper we propose a new model to detect multiple objects of various intensities in images having maximum, minimum, or middle-intensity background by evolving only one level set function. In this model, a new signed pressure force function based on novel generalized averages is used for segmentation of images with maximum or minimum intensity background. For images with middle-intensity backgrounds, which are indeed challenging for 2-phase models, we propose a new product generalized signed pressure force function. Finally, to give experimental and qualitative evidence, our model is tested on both synthetic and real images with the Jaccard similarity index. The experimental and qualitative results reveal that the proposed method is efficient in both global and selective segmentation. Our new model is also tested on color images and the results are compared with the state-of-the-art models.