Multi-Class Cancer Classification via Partial Least Squares with Gene Expression Profiles
We describe a hybrid data-representation and rendering technique for visualizing large-scale particle data generated from numerical modeling of beam dynamics. The basis of the technique is mixing volume rendering and point rendering according to particle density distribution, visibility, and the users instruction. A hierarchical representation of the data is created on a parallel computer, allowing real-time partitioning into high-density areas for volume rendering, and low-density areas for point rendering. This allows the beam to be interactively visualized while preserving the fine structure usually visible only with slow pointbased rendering techniques.