中文摘要：

提出了一种新的硬件加速自适应EWA（椭圆加权平均）Splatting算法框架，可同时适用于三维体数据和点模型．算法将高斯重建核与低通图像滤波核结合，得到反走样、无模糊的高质量图像．提出一种高效的自适应滤波方法，减少了高质量EWA Splatting的计算量．提出了自适应体EWA Splatting的3种数据存储模式和一系列高级特性，其中包括交互式分类、体一面混合绘制策略和自适应浮点累加．展示了如何在可编程图形处理单元（GPU）中计算体数据和点模型数据的EWA Splat基元．实验表明，文中的方法在一台普通微机上每秒可绘制1500万～2000万个基元，达到较高的图像质量与交互的绘制速度．

英文摘要：

This paper presents a novel framework for hardware-accelerated adaptive EWA （elliptical weighted average） Splatting. EWA Splatting combines a Gaussian reconstruction kernel with a low-pass image filter for high image quality without aliasing artifacts or excessive blurring. This paper introduces an efficient adaptive filtering scheme to reduce the computational cost of high quality EWA Splatting, and shows how to compute the EWA Splat primitives for volume data and for point-sampled surface data on modern graphics processing units （GPUs）. To accelerate the rendering, the splat geometry and data attributes are assembled locally in video memory. For adaptive EWA volume Splatting, it proposes three data storage modes and several advanced features including interactive classification, hybrid surface-volume rendering and adaptive floatingpoint accumulation. The current implementation renders 15-20 millions primitives in a consumer PC. Several results for rectilinear volume data and point-sampled surfaces demonstrate the high image quality and interactive rendering speed of the proposed approach.