GPU-based reconstruction and display for 4D ultrasound data

Due to the required computational effort of 4D ultrasound imaging, such systems depend on low complexity techniques like nearest neighbor interpolation, which affects volume quality. Moreover, more accurate techniques like normalized convolution, backward trilinear interpolation, and forward spherical and ellipsoidal Gaussian kernel, are avoided in real-time imaging because of the tight reconstruction time. The goal of this work is to utilize recent commercial graphics hardware technology of graphics processing unit (GPU) to speed up the reconstruction time while increasing the quality of displayed volume.