A Two-Level Differential Volume Rendering Method for Time-Varying-Volume-Data



Shih-Kuan Liao, Yeh-Ching Chung[1], Jim Z.C. Lai


Department of Information Engineering

Feng Chia University, Taichung, Taiwan 407, ROC

Tel: 886-4-24517250 x3765

Fax: 886-4-24516101

Email: {skliao, ychung, lai}@fcu.edu.tw






The differential volume rendering method is a ray casting based method for time-varying-volume-data.  In the differential volume rendering method, the changed fractions of volume data between consecutive time steps are extracted to form differential files.  Based on the differential files, only the changed pixels, instead of all the pixels in the image, are updated by casting new rays at the positions in each time step.  The main overhead of the differential volume rendering method is to determine the changed pixel positions before casting new rays for the changed pixels.  In this paper, we propose a two-level differential volume rendering method, which is a modified differential volume rendering method with faster determination of the changed pixel positions.  In the proposed method, the determination of the changed pixel positions is accelerated by the aid of second-order-difference.  Since voxels in two consecutive differential files may partially overlap in the space, the computation spent on determining the changed pixel positions due to the overlapped area is redundant.  We use this property to extract the difference of changed voxel positions between consecutive differential files to form the second-order-difference.  Based on the second-order-difference, the changed pixel positions can be determined efficiently.  The experimental results show that the proposed method outperforms the differential volume rendering method for all test datasets.


Keywords: Ray casting, differential volume rendering, time-varying-volume-data, flow animation, CFD.


[1] The corresponding author.