Liver segmentation from CT image combined with deformation and triangle subdivision model

As the traditional deformable model is easy to fall into local optimum, high precision segmentation for the liver depression area is difficult. In order to solve this problem, a novel algorithm based on deformation and triangle subdivision model is proposed for the liver segmentation from CT images. First, a simplex mesh model is established to represent the initial boundary of the liver. Second, a new set of internal force and constraint model are constructed based on the relationship between the vertex and its three neighborhoods vertices. Third, the external force and the constraint method are established by using the combination of the balloon force model and the Gabor feature of images to push the model more quickly approaching the liver boundary under the action of internal force and external fore. Forth, in the process of the model deformation, new vertices can be inserted to ensure the smoothness of the model and the accuracy of segmentation results for sag area of liver. Finally, the data provided by international conference on medical image computing and computer assisted Intervention are used to investigate segmentation results. The experimental results show that the proposed method for liver segmentation has better applicability and robustness, as well as gets higher division accuracy.