Segmentation MC algorithm-based method for ultrasound image 3-D reconstruction

The traditional marching cubes (MC) algorithm has several shortcomings, including that it has to process all the voxels in the space so the execution is not efficient and the segmentation threshold value is difficult to determine. A segmentation-based MC algorithm for three-dimensional reconstruction surface rendering of ultrasound images was developed to improve the speed and image quality. The algorithm uses divisional images as the input, with segmentation results to construct an iso-surface. The result is faster with better quality of 3-D reconstruction. Ultrasound images of liver tumors and the surrounding tissues and organs (blood vessels) were reconstructed with the visualization toolkit (VTK). Tests show that the algorithm requires less than 20 s and is very precise, sufficiently for clinical requirements. The algorithm has been used in successful clinical applications for minimally invasive visualization of liver cancers.