GPU-assisted energy asynchronous diffusion parallel computing model for soft tissue deformation simulation

Abstract

Soft tissue deformation simulation is a key technology of virtual surgical simulation. In this work, we present a graphics processing unit (GPU)-assisted energy asynchronous diffusion parallel computing model which is stable and fast in processing complex models, especially concave surface models. We adopt hexahedral voxels to represent the physical model of soft tissue to improve the visual realistic quality and computing efficiency of deformation simulation. We also adopt the concept of free boundary to simulate soft tissue geometric characteristics more precisely during the deformation process and introduce asynchronous diffusion by using the mechanical energy of mass points to achieve realistic soft tissue deformation effects. In order to meet the requirement of real-time surgery simulation, we accelerate the soft tissue deformation by using OpenCL (Open Computing Language) and optimize the parallel computing process in several means. Experimental results have shown that the GPU-assisted energy asynchronous diffusion parallel computing model for soft tissue deformation simulation implements satisfactory effects on deformation in visual realistic and real-time quality.

A fast mass spring model solver for high-resolution elastic objects

“…where 3D elastic object simulation plays a central role. [1][2][3] Based on this, efficient and visually realistic deformation effects of high-resolution geometry models of 3D elastic objects are needed in many scenes. By far, the most widely used methods to simulate elastic objects are through geometric models, finite element models and mass spring models (MSMs).…”