Liver Segmentation Method in CT Image Based on Adaptive Surface Deformation Model

Abstract

The invention provides a self-adaptive surface deformation model based CT (Computed Tomography) image liver segmentation method and a convenient tool is provided for the clinical diagnosis of liver diseases. The self-adaptive surface deformation model based CT image liver segmentation method comprises step 1, performing preprocessing on an image by an anisotropic filtering method to obtain an initial boundary image; step 2, describing the initial contour of the liver through a DSM (Deformable Simplex Model); step 3, calculating the internal force of the model according to the relation between a vertex of the model and the neighborhood of the vertex and calculating the external force of the model through the gradient and the boundary of the original image; step 4, building a constraint model of the internal force and the external force of the model; step 5, performing self-adaptive decomposition of a triangular mesh; step 6, setting the number of iterations and approximating a target area under the driving of the internal force, the external force and the balloon force; step 7, obtaining an accurate segmentation result.

Description

technical field [0001] The invention relates to a CT image liver segmentation method based on an adaptive surface deformation model, belonging to the image scanning field of liver segmentation. Background technique [0002] The liver is the largest metabolic organ in the human body, the synthesis, decomposition, transformation and storage center of the substances and nutrients needed by the human body. Liver segmentation is an important basis for computer-aided diagnosis and treatment, tumor resection, surgical planning and living donor transplantation. The result of liver segmentation directly affects the accuracy of liver three-dimensional reconstruction, and affects doctors' diagnosis and treatment of liver lesions. Therefore, the research on liver segmentation method not only has important academic significance but also has important clinical application value. [0003] However, in computed tomography (CT) images, the gray value of the liver tissue is often similar to ...

AI Technical Summary

Problems solved by technology

[0004] 1. The initial position of the deformable model seriously affects the final segmentation result. If the model initialization is too small, it is easy to cause the model to reach local equilibrium or shrink to a point.

At this time, the liver tissue cannot be effectively segmented

[0005] 2. If the model initialization is too large, it is easy to cause the deformation model to overflow the segmentation area, resulting in over-segmentation

[0006] 3. Due to the influence of the smooth force of the deformable model, it cannot effectively enter the slender and sharp corners of the liver, which seriously affects the segmentation accuracy

Images