Abstract
An acceleration technique using bilinear and simplified trilinear interpolation which reduces the amount of interpolation computation during the ray sampling procedure of the volume ray casting algorithm is described. It has several characteristics: It is applicable to both perspective projection and parallel projection; it uses almost no extra space and processing power; there is no noticeable sacrifice of image quality; and more importantly, it can be easily applied to existing ray casting optimizations to further improve their efficiency. In addition, an adaptive depthsampling optimization has been developed based on our technique, which serves as an effective approach to meet the frame rate demand of time-critical visualization environment. Experimental results show that a rendering rate of more than ten frames per second has been achieved on a 16-processor SGI Power Challenge.