| Image segmentation is the one of the key technologies in field of computer vision and image processing. Especially in recent years, with the rapid development of computer technology, it has attracted the attention of many researchers. Digital Image segmentation is applied more widely in clinical medicine. Scholars have proposed a lot of models: the more prominent is the active contour model Type. The active contour model(Snake),originally presented by Kass in 1987,has been widely used in image processing fields, such as image segmentation ,edge detection and motion tracking.Based on the traditional Snake model image segmentation, this thesis further studies the medical image segmentation based on the GVF Snake Model. It preferably solves the initialization problem and concavity problem which is hard to deal with by the traditional snake model. However, in some image segmentation, GVF Snake still can not convergence to the target contour well. In this paper, Perona-Malik is introduced before image segmentation, it enhances image edge while abates noise, which makes the position accurate. Then some of the shortcomings of GVF Snake are modified. An adaptive image structure GVF Snake model is proposed. This model takes into account the specific structure of the edge of the image information, so it has better convergence results than GVF Snake.Three-dimensional image sequence segmentation is based on two-dimensional image segmentation, so two-dimensional segmentation is particularly important. Based on the characteristics of 3D images, i.e. the surface is smooth and the deformation of the contour is continuous in 3D space,the adaptive image structure GVF Snake model is applied in 3D images segmentation. Experiments show that the adaptive image structure GVF Snake model can achieve better segmentation results in two-dimensional and three-dimensional image segmentation. It also can remarkably improve the accuracy of the segmentation in comparison with the GVF Snake. |
PDF Full Text Request