Three-dimensional visualization of cardiac volume data based on GPU-based raycasting algorithm

Abstract

Imaging and visualization of heart play an important role in cardiac disease diagnosis and therapy planning. To visualize the segmented cardiac volume datasets, an approach based on Graphics Processing Unit( GPU) acceleration raycasting algorithm was proposed to reconstruct high-quality image. Firstly, the transfer function was designed by using statistical information of the cardiac volume datasets, which was used for increasing opacity of the subtle tissues. Secondly, the sampling step was adaptively adjusted based on the gradient, which was utilized to increase sampling frequency of the borderline tissues. Finally, the improved Blinn-Phong illumination model with multiple lamp-houses was also employed for enhancing the visualization effect. The experimental results show that the proposed method can meet the requirements of realtime rendering and obtain high-quality volume rendering results. Cardiac subtle tissues are rendered clearly such as valves and coronary artery veins.