-
Notifications
You must be signed in to change notification settings - Fork 30
Open
Description
Hi, i have a query point set and a second point set,
for each points in the query, i am asking the closest points of the second set, in the radius of 1.0
i can see i am limited to 70 neighbours, but what if they are more :
- are the 70 points classified by euclidian distance ?
- are they the 70 closest, or 70 points randomly taken ?
what would be the best strategy to recover ALL Neighbors from the radius ( 50k ? ) and => classify them by distance for the given radius => in order to have the 70 closest points ? ( CUDA_MAX_NEIGHBORS )
i understand there is a theory of sorting the points by "space-filling Z curve" how does it differ from euclidian distance ?
by reading "FAST FIXED-RADIUS NEAREST NEIGHBORS: INTERACTIVE MILLION-PARTICLE FLUIDS"
i understand there a way of "Finding all neighbors in a fixed radius R" or is there a limitation ?
thank you !
Metadata
Metadata
Assignees
Labels
No labels