Multiphase Image Segmentation Variational Level Set Method

Image segmentation is an important research aspect in computer vision.And the goal of it is to separate different regions or detect objects based on gray,texture,color,optic and so on.As the complexity of the object itself and the requirements of multi-objects to detect in images,traditional method of single level set segmentation cannot satisfy the situation, and it is mostly limited in the 2D field.So,we need to do profound research work for multiphase image segmentation and extend the level set method from 2D into 3D field for higher requirements.Several aspects are introduced in this paper as follows:Firstly,the basic theory on curve evolution,level set method and Mumford-Shah Model are systematically researched and their application in image segmentation field is discussed. Secondly,general regions representation model is brought out based on the region competition strategy of Abdol-Reza Mansouri without overlapping and vacuum problems. For the piecewise constant case,different general multiphase image segmentation models are established based on the Paragios's geodesic active region model and used in different probability distribution,for instance,MRI image fits Gauss Probability Distribution, Ultrasound image fits Rayleigh Probability Distribution.Thirdly,the model is extended into piecewise smooth case.The smooth data term is kept in the model and the power is changed into one power instead of tow power.It is better to keep edge in the progress of segmentation.Fourthly,the variational level set method is extended from 2D into 3D field and the same region compete strategy is also used in the 3D image segmentation.And different models are established based on different probability distribution just like the 2D models and the progress of segmentation is also the progress of 3D reconstruction. Furthermore,all the models mentioned above are implemented,including piecewise constant/smooth multiphase image segmentation for 2D images and the piecewise constant multiphase image segmentation for 3D images.Many experiments denote that we get the satisfying results using the corresponding image segmentation models.Lastly,future research directions are put forward according to the problems and limitation occurred during my research work.