Semi-Automatic Generation of Stream Surfaces Via Sketching

Jun Tao,Chaoli Wang
DOI: https://doi.org/10.1109/tvcg.2017.2750681
IF: 5.2
2018-01-01
IEEE Transactions on Visualization and Computer Graphics
Abstract:We present a semi-automatic approach for stream surface generation. Our approach is based on the conjecture that good seeding curves can be inferred from a set of streamlines. Given a set of densely traced streamlines over the flow field, we design a sketch-based interface that allows users to describe their perceived flow patterns through drawing simple strokes directly on top of the streamline visualization results. Based on the 2D stroke, we identify a 3D seeding curve and generate a stream surface that captures the flow pattern of streamlines at the outermost layer. Then, we remove the streamlines whose patterns are covered by the stream surface. Repeating this process, users can peel the flow by replacing the streamlines with customized surfaces layer by layer. Furthermore, we propose an optimization scheme to identify the optimal seeding curve in the neighborhood of an original seeding curve based on surface quality measures. To support interactive optimization, we design a parallel surface quality estimation strategy that estimates the quality of a seeding curve without generating the surface. Our sketch-based interface leverages an intuitive painting metaphor which most users are familiar with. We present results using multiple data sets to show the effectiveness of our approach.
What problem does this paper attempt to address?