Research On Edge Extraction In MRI Brain Images Based On Snake

Edge extraction is a basic technique in the image processing and computer vision, and it has become the foundation of image segmentation, object recognition and image compression. The major goal of this paper is to systematically develop and apply the technique of Snake for edge extraction in MRI brain images.This paper summarizes the methods of 3D reconstruction in MRI, and principally researches on the steps of 3D surface reconstruction in order to confirm that edge extraction is in the key status of 3D reconstruction.Active contour model(Snake)is a flexible tool of using various of restricted conditions, but this model has a strong pertinence and weak currency. In order to solve these faults, an algorithm based on the improved Snake is proposed. The model not only precisely extracts the edge of the protruding object in images, but can also near to the edge of the concave object; it introduces exterior sanction energy which can adaptively change its value to enlarge suction, and it makes control points quickly converge to the real contour without depending on the initial contour.This paper designs edge extraction technology composed of Snake and other segmentation methods. Snake is combined with equal-distance threshold segmentation and also integrated in the generalized fuzzy operator to extract head, experiment shows that the above methods are available. It adopts the semi-automation method of user intervention, and can quickly near to the tumor edge, the effect is satisfaction. The method is consisted of diffusion, precognition and Snake for brain edge, experiment proves that it has a good automatic and robust character. This paper is aimed at the continuity of slices, and proposes a 3D edge extraction method consisted of edge extraction, edge transfer, and combination all the edges. An edge transferring methods is build, which is successfully combined with the image matching and can increase the precise of transferring the edge from one slice to others. Experiments shows that this algorithm can obtain the edges of the desired object from MRI volume images quickly and reliably, and it has practical value in the medical image analysis.