PointCast
In a 3D particle cloud on a 2D screen, a single click sends a ray into the data; the system reads the density along the ray and selects the first real cluster it passes through, skipping faint ones.
How it works
The density field is sampled along the ray through the clicked point, and stretches above 10% of the maximum are treated as clusters. The nearest one is selected unless its accumulated density is under 10% of the strongest, in which case the strongest is taken. The selection is grown from that point with a threshold of 20% of its peak density, and lowering the threshold afterwards reveals how the cluster connects to others. Participants preferred it most, especially for partly occluded clusters, since only a small visible part is needed. Compared on a pen tablet (20 participants) with cylinder selection and CloudLasso: the CAST techniques took 16 to 23 s against 43 s for CloudLasso and 88 s for cylinders, with the same accuracy.
Papers
- Lingyun Yu, Konstantinos Efstathiou, Petra Isenberg, Tobias Isenberg (2016). CAST: Effective and Efficient User Interaction for Context-Aware Selection in 3D Particle Clouds. IEEE Transactions on Visualization and Computer Graphics 22(1), 886-895 doi:10.1109/TVCG.2015.2467202