高级检索

基于GPU的多分辨率体数据重构和渲染

GPU Based Multi-resolution Volume Data Reconstruction and Rendering

  • 摘要: 基于小波变换的多分辨率压缩算法能够获得很高的压缩比,因而被广泛地用于压缩体数据.针对这种压缩策略,研究基于GPU的数据重构的方法,可以只从CPU向GPU传输少量的压缩数据,从而提高数据传输效率.因为好的数据结构是实现基于GPU的重构算法的关键,所以文中提出适合使用矩形纹理表示的数据结构——NestedTileboard;然后给出基于该数据结构在GPU上实现多分辨率重构的方法,使用NestedTileboard保存中间数据及重构结果;还提出了基于NestedTileboard的多分辨率体绘制方法,直接对重构数据进行体绘制,从而实现数据重构和体绘制的无缝连接.

     

    Abstract: Wavelet transform and multi-resolution based compression algorithm can be used to achieve high compression ratio,so it is widely applied to volume data compression.On the compression algorithm,the paper investigates the method directed to GPU-based data reconstruction,with purpose of transferring only the compressed data from CPU to GPU.Because an advantageous data structure is crucial in GPU-based data reconstruction,we propose such a data structure named Nested-Tileboard,which can be represented as rectangle texture.A multi-resolution data construction method is accordingly proposed based on Nested Tileboard to store intermediate data and reconstruction result.The paper also presents Nested-Tileboard volume rendering method,to allow the reconstructed data rendered directly,and the volume data reconstruction and rendering processed seamlessly.

     

/

返回文章
返回