Model Based Visualization Of Four-Dimensional Data

Yh Huang,Lz Ma,Zd Jiang
2001-01-01
Abstract:The visualization of three-dimensional (3D) scalar data field is one of the important research topics in scientific data visualization. In the paper we study model based four-dimensional (4D) data visualization techniques due to the fact that 3D scalar data field is one instance of 4D data fields. Instead of rendering the 4D data directly by projecting them onto 3D subspaces, we reconstruct a continuous 4D super-surface model from the given data field and employ 4D rendering algorithms to display it. Interpolating or fitting approach is adopted to reconstruct the continuous 4D super-surface model according to the categories of the given data field. When we have no intuitive knowledge of the given data field, 4D-Bezier or 4D-Bspline super-surface model is adopted to interpolate it while the Least Square fitting model method is employed in the case that we have enough information of the given data field. The reconstructed continuous super-surface is rendered by 4D raytracing and slice rendering methods. The model result (a continuous representation of the 4D data field) is reusable, analyzable and unique. Experimental results on several 4D mathematical data fields and one measured data field from real world show that the proposed method can reveal the whole features of the given data fields effectively and vividly.
What problem does this paper attempt to address?