Research Of Medical Three-dimensional Reconstruction Based On GPU And Energy Function

Scientific visualization is a new field which was proposed and developed in1980s.With the development of computer application technology, scientific visualization hasbeen applied to medical field. Thus, the medical visualization technology is born, and3Ddata field visualization is the core content. The medical data visualization technology is a3D model rebuilded and demonstrated technology, which is based on2D information fromCT or MRI. Through medical visualization, the medical staff can observe the nidus fromdifferent perspectives, and make scientific and efficient diagnoses.First, with a wide knowledge of3D data visualization, iso-surface reconstructiontechnology is deeply researched, especially the face drawing method of iso-surfacereconstruction technology. Marching Cubes algorithm is the most popular in isosurfacereconstruction technology area. Through analysis of Marching Cubes algorithm, MarchingCubes algorithm is improved in aspect of consuming long time and large calculatingamount. Regional boundary growing algorithm can resolve this problem well.It will bemore fast with the GPU.Second, the reconstructed three-dimension model is optimized. The size oftrianglefaces is different, which is consist of three-dimension model, made thethree-dimension model not enough delicate. Thus, if the area of triangleface greater thanthe threshold F, It will be divided with the center of gravity. Then, a energy function isproposed. The new coordinates of the triangleface’ vertexes can be figured out throughextremum principle, thus making the3D model approximates the real object.Finally, with the CT data of skeleton provided, reconstruct the data with the improvedalgorithm and meanwhile optimize the3D model through optimization algorithm.Consequently, the experiment result proves that the algorithm is feasible and effective.