Research And Application Of Transfer Function Specification In Direct Volume Rendering

Volume rendering is a key technology for the visualization of three-dimensional dataset. Transfer function is used to specify the relationships between volume data and optical properties in the process of volume rendering. Specification of transfer function plays an important role in the rendering quality. Unfortunately, specifying a good transfer function is a difficult and tedious task, thus it has been referred as one of the top ten problems in volume visualization.The application of hierarchical clustering into transfer function specification is shown to be effective. However, with the limited capabilities on computing and storage, it is very difficult for a single processor to accomplish cluster analysis of the large scale 3D volume data within an effective period. In this paper, a parallel hierarchical clustering algorithm, which is combined with the LH histogram, is designed in order to instruct effectively the specification of transfer function. Compared with the serial algorithm, our parallel algorithm is able to accomplish the clustering analysis of the large scale 3D volume data within a reasonable period. The images with high quality are rendered finally.On the other hand, supposing the process of finding a suitable transfer function to be a parameter optimization problem, the global optimization approach is a solution. So far genetic algorithm has been applied into the research on transfer function specification. In this paper, the particle swarm optimization is applied into this field with two different kinds of evolutionary approaches of automatic and interactive. Besides, the user-based speed update strategy in interactive approach is also proposed. The basic and improved particle swarm algorithms, which are capable to provide user with satisfying volume rendering images much faster, have better performance than genetic algorithm.The system of seismic data visualization provides users with a two-level tool for transfer function specification by combining the particle swarm optimization algorithm, manually adjusting transfer function and multi-resolution displaying together. With this tool, which is the combination of automaticity and interactivity, the user can get real-time feedback of the rendering result, which can make geophysicists explore the seismic data more expediently and flexibly, and enhance the accuracy and success rate of seismic exploration.