Medical Image Segmentation Based On Shape Learning And Curve Evolution

The development of the technology of medical image has provided a platform for the researchers of medical image. The segmentation of anatomical structures from medical images is a very important step in the medical image analysis and processing. It is the foundation work of the biomedical image, clinic diagnose and pathology analysis. The level set curve evolution segmentation models are now the very popular method of the medical image segmentation. They can cope with the change of topological structure and have been widely applied in the medical images. However, meantime they also have some limitations. First, shape prior knowledge is seldom used in the models. Then, Intensity inhomogeneity often occurs in real images from different modalities, for medical images, intensity inhomogeneity is usually due to technical limitations or artifacts introduced by the object being imaged. In such situation, the traditional level set curve evolution models cannot segment medical images correctly.In order to introduce the shape prior knowledge to the level set curve evolution models, this paper presents a new level set model which incorporates the shape information to restrict the curve evolution. The new method is based on Chan-Vese model, we first apply the statistical shape learning to the target object which needs to be segmented, then build a shape energy function and add it to the model energy function, and then minimize this total energy function, and finally get the correct segmentation results. We apply this new method to the brain MRI images, and segment the brain skins. The experiment results demonstrate that the new method has the very good segmentation results.Intensity inhomogeneities often occur in real-world images and may cause considerable difficulties in image segmentation. In order to overcome the difficulties caused by intensity inhomogeneities, this paper presents a novel level set method for image segmentation. Gray-level moments are used to estimate two fitting functions that approximate local intensities on the two sides of object boundaries, which are then incorporated into a variational level set framework. Energy functional is defined on a contour, which characterizes the approximation of local intensities on the two sides of the contour by the two fitting functions. This energy can be minimized when the contour is on the object boundary. Thus, image segmentation is performed by minimizing this energy functional. A desirable feature of our method is that it is not sensitive to the contour initialization and is able to segment images with intensity inhomogeneity. Only a small number of iterations are needed to obtain the final result, which makes our method more efficient than previous level set methods.