Multipass GPU surface rendering in 4D ultrasound

Abstract

4D ultrasound imaging extends the real-time ca-pability of ultrasound to display a real-time volume, which can then be explored by the sonographer. Surface Rendering is one of the common modes used to display the 3D datasets. Surface shading is required to visualize the surface and en-hances the surface contrast. Here, we show a surface rendering implementation based on three passes. In the first pass, ray casting is used with edge detection algorithm and then the edges are stored to viewing plan distance in the z-buffer of the 2D rendered image. In the second pass, filtration of the z-buffer is performed. This stage is so critical because the fuzzy nature of the ultrasound dataset then shade the 2D rendered image using image space shading. Finally, the third pass applies additional image processing mainly image smoothing on the rendered 2D image. The new method is applied to render volumes acquired on a research system and quality and computation time results show potential for clinical utility.