3D Reconstruction Based on Image Segmentation

Qiang Weizhe,Song Guanghua,Zheng Yao
DOI: https://doi.org/10.3321/j.issn:1002-8331.2005.36.025
2005-01-01
Abstract:3D reconstruction is important for medical image visualization and treatment plan system.This paper presents a method of 3D reconstruction based on image segmentation,which is different from traditional Marching Cubes algorithm.The method consists of image segmentation,isosurface extraction and mesh simplification.Firstly,the images are segmented to binary images,then the isosurface is extracted by means of Marching Cubes,lastly the mesh representation is simplified.Experimental results are given to show that the present method improves the performance of Marching Cubes,enhances the result of reconstruction and realizes the real-time rendering and interaction of large geometry models.
What problem does this paper attempt to address?