Interactive GPU-Based Volume Rendering for Medical Image

As we know that the 3D medical images are usually in large scalar, the traditional volume rendering method which running on CPU can not achieve interactive rendering speed. Recently, GPU-accelerated direct volume rendering has positioned itself as an efficient tool for the display and visual analysis of volume data. To facilitate the usage of well developed hardware resource, a graphics processing unit (GPU)-based volume ray-casting algorithm is proposed in this paper. The proposed algorithm implement ray casting operation completely in GPU. The algorithm re-samples volume data, represented as a stack of 3D texture, onto a sampling surface. The ray-casting algorithm performs in fragment shaders. Additionally we apply early ray termination. At last we demonstrate the effectiveness of our algorithm with several medical volume data. It proved that the proposed algorithm can perform an interactive rate even for direct volume rendering while keeping the high image quality.