Method and system for partitioning construction in CT image

Abstract

The invention provides a method and system for partitioning a construction in a CT image. The method includes the steps that the opacity of the CT image is adjusted, and a construction structure in the CT image is obtained; the construction to be partitioned is determined in the obtained construction structure, interactive partition is conducted on one layer of the CT image of the construction to be partitioned, the processes are that firstly, the gradient vector is used for constructing an energy cost function of the CT image under the opacity, a seed point is selected at the position close to the construction to be partitioned, the optimal path, generated according to the energy cost function, between a current point and the seed point is the outline of the image of one layer of the CT image of the construction to be partitioned, secondly, the outline of a layer adjacent to the layer where the outline is located is generated, thirdly, outlines of the images of all layers in the CT image of the construction to be partitioned are generated according to the iteration method, fourthly, the outlines of the images of all the layers in the CT image of the construction to be partitioned are combined to obtain the three-dimensional construction to be partitioned. By means of the method and system for partitioning the construction in the CT image, interactive partition can be conducted on the three-dimensional CT image rapidly and accurately.

Description

technical field [0001] The present invention relates to the technical field of image segmentation, and more specifically, to a method and system for segmenting tissues in CT images. Background technique [0002] The three-dimensional segmentation of medical images can provide a reliable basis for clinical diagnosis and pathological research. However, images often have complexity and ambiguity. At present, there is no standard method that can solve all segmentation problems. Therefore, many segmentation tasks cannot be done automatically by computer, and manual segmentation has a problem of heavy workload. Because of this, people have proposed some methods combining manual interaction and computer automatic positioning to realize the rapid positioning method of the target contour. This fast positioning method can be roughly divided into two categories: [0003] One is the Snake algorithm, which defines an energy curve, which is jointly determined by internal energy and exter...

AI Technical Summary

Problems solved by technology

On the one hand, the energy value of the short path between the seed points is more dominant than the energy value of the long path, so the point on the shortest path cannot be guaranteed to be an edge point, so the contour cannot be tracked correctly, which makes the contour of the tissue Segmentation accuracy is not high; on the other hand, when using this intersection contour segmentation method for 3D segmentation, it is first necessary to perform interactive contour segmentation in two key layers, and then use an interpolation algorithm to generate intermediate layer contours between key layers , but the traditional interpolation algorithm has the disadvantages of noise sensitivity, discontinuity, and large time overhead, which further affects the effect of image segmentation

Images