基于GPU的医学图像快速体切割算法

Fast Volume Clipping of Medical Images on GPU

目的:在体绘制过程中,有许多重要的、医生感兴趣的细节信息隐藏在数据场内部,在进行显示时,这部分信息很容易被其他组织或器官遮挡,无法显示在重建图像中,为了给医生提供全面、直观和准确的诊断信息,本文提出了一种基于GPU加速的体切割算法。方法:通过将切割算法和基于GPU的光线投射算法结合,实现体数据的快速切割。本文在基于GPU加速的医学图像快速体绘制的基础上,将剖面的空间信息传入着色器,然后通过比较体数据的空间坐标与剖面位置的关系来决定体数据的取舍。该方法不同于以往基于深度模板信息的体切割,在定义好切割平面后,可从任意角度对保留下来的有效体数据的重建结果进行观察。结果:该方法能够精确地按照用户定义的形状对体数据进行切割,并且由于使用了硬件的加速功能,该方法可以达到实时交互的速度。结论:该方法能够满足医学影像可视化的实时交互要求,在手术模拟等临床技术中有广泛应用。

Objective： When the region of interest is somewhere in the middle of the volumetric data, it will be occluded by other regions of the volumetric data and will be hard to see in the reconstruction of volume rendering. In order to provide doctor with comprehensive, visual and accurate diagnostic information, fastvolume clipping of medical images on GPU is proposed in the paper. Methods： The rapid volume clipping is achieved by combining the clipping algorithm with the fast volume rendering algorithm based on GPU. The technique is based on fast volume rendering of medical images on GPU. It works as follows： The space information of the section plane is sent to the shader. Then compare the position of the volume data sets with the section plane to decide whether the data is clipped or not. Our algorithm is different from the algorithm that clipping based on depth stencil. We can see the result that reconstructed from the data sets that remained from any position as long as the section plane is defined. Results： According to the shape the users defined, our algorithm can decide which parts of the volume have to be clipped. Because of using the general graphics hardware acceleration, the proposed method achieves interactive display rate. Conclusions： The experiment shows that the algorithm is significantly fast and can meet the requirements of real-time interactive rendering on the premise of ensuring imaging quality and can be used in surgery simulation and so on.