“Similarity metrics for bounding volumes” by Patel
Conference:
Type(s):
Title:
- Similarity metrics for bounding volumes
Presenter(s)/Author(s):
Abstract:
Testing the similarity of bounding volumes is a common operation in geometry processing. Often, the intersection or union of two bounding volumes is calculated as an intermediate step. These intermediate volumes are then measured to produce a comparison metric. Common metrics include distance, surface area, volume and overlap. Here, we introduce new variations of these metrics, using a new intermediate calculation inspired by interval arithmetic. We argue that the new technique offers better precision than calculating intersections or unions of bounding volumes because more detail is available over the entire range of inputs.
References:
1. Aggarwal, C., et al. 2001. On the Surprising Behavior of Distance Metrics in High Dimensional Spaces. Proc. 8th International Conference on Database Theory. 420–434.
2. Henrikson, Jeff. 1999. Completeness and Total Boundedness of the Hausdorff Metric. MIT Undergraduate Journal of Mathematics, 1. 69–80.
3. Levin, V. I. 2004. Ordering of Intervals and Optimization Problems with Interval Parameters. Cybernetics and Systems Analysis, 40, 3. 316–324.
