| Volume rendering of medical image is a heat and classical issue in the field of visualization. It makes best use of the volume data of CT or MRI to render3D image of any angle needed, which brings much convenience to observe the physical structure for doctor. Thanks to the interactive procedure, the doctor can get multi-angle and multi-level means of observation and analysis.But huge data demand higher requirement for visualization. How to realize real-time rendering for large-scale data becomes a closed question. Thus, the compression of volume data becomes a active study field. Under the support of "National Digital Medical Image Equipment&Engineer Technology Research Center", this paper undertakes the research of acceleration and compression of volume rendering algorithm are as follows:Firstly, the analysis and the read of DICOM document are implemented according to the format and data coding method of DICOM medical image. Consequently, the DICOM image can be directly used during the period of volume rendering.Secondly, all main algorithms about volume rendering are studied. In term of the require of real-time rendering in practice, the Proximity Clouds algorithm is performed to accelerate volume rendering, which increase the rendering speed by70%. Furthermore, the improvement of Grads measurement is proposed, which increase the rendering speed again in the ray casting algorithm based on GPU.Finally, the incremental compression algorithm is proposed to compress volume data in ray casting algorithm. Compare to the traditional VQ and wavelet compression algorithm, the incremental compression algorithm is lossless, no-decompressed, real-time rendering, and compression rate can reach1:2(such as Heart data reaching1:2.5; Head data reaching1:1.8). But the pretreatment time is long. Moreover, we also realize the dynamic loading of volume data algorithm which makes it possible to load large scale data to Graphics card of limited capacity layered. |
PDF Full Text Request