P‐2.16: Research on three‐dimensional reconstruction based on VTK

Abstract

Three‐dimensional reconstruction of medical tomographic images is a technological way to provides a 3D model of biological tissue. As it not only provides a 3D model, but also shows the spatial location of this model, it received widespread attention. By human‐computer interaction, doctors can easily conduct disease analysis and surgical planning.This paper introduces a novel three‐dimensional reconstruction technology of magnetic resonance images, which deals with the brightness, contrast adjustment, and three‐dimensional reconstruction of Dicom images. This paper uses the Visualization Toolkit VTK to implement Dicom 3D image reconstruction. VTK has a large number of callable functions built in and encapsulates some common visualization algorithms. When visualizing medical data, VTK can directly call the related functions. VTK three‐dimensional reconstruction can help people intuitively see the contents of a 3D image. Displaying two‐ dimensional images is easy, but displaying 3D images is not. Previously, because of the limited technical means, even if the data of the 3D image is obtained, it can only be displayed as a slice of the 2D image, which is not intuitive. With the development of computer science and technology, the use of computer graphics technology to model and render 3D objects in real‐time has become the mainstream method of object display, while the past two‐dimensional display method has been gradually phased out.To describe the process of 3D modeling more accurately and intuitively, this paper focuses on a reconstruction technique, polygon drawing. Polygon drawing uses segmentation technology to perform contour recognition, extraction, and other operations on a series of two‐dimensional images, and ultimately restores the three‐dimensional model of the detected object and displays it as a surface.First, the original data is a discrete three‐dimensional spatial rule data field. The images generated by CT and MRI of tomography scanners belong to this data type, that is, the DICOM type image used in this paper. Second, the boundary between the content part and the background part in the three‐dimensional image is found in the unit of volume, and triangles are extracted to fit the boundary. Then, all the voxels are traversed to find out which triangles are finally assembled to form a triangular mesh of the surface of the real point in the image. Finally, a triangle mesh model is built to render the model, and the rendered model is put into the window to interact with users.