For beta there where a k-means addon that could do clustering.
Is there something like that in gamma, I am looking around and there is lots of advanced recognition, but no simple clustering or am I just looking in the wrong places?
Sune
For beta there where a k-means addon that could do clustering.
Is there something like that in gamma, I am looking around and there is lots of advanced recognition, but no simple clustering or am I just looking in the wrong places?
Sune
Not sure what you’re after exactly, but you might wanna have a look at VL.DBSCAN if you’re interesting in 2d clustering.
The current nuget version only handles 2d, but @antokhio recently made a PR I still have to merge that might (if I got that right) allow to cluster more diverse things :-)
Cheers
@sebescudie The PR I made, is just an extension to allow usage of something else then Spread<Vector2>
for clustering.
But yes it works quite nicely with simple Spread<Vector2>
. The performance is’t really that great tho, I think it’s ok up to 5 or 10 k points (but not sure). As I remember (but might be wrong), this algo is’t great for large datasets, there is some other algos for faster clustering.
I will be using it on reduced kinect point cloud. The way it is set up, discarding the Z component is fine.
hello @sunep , specifically for clustering and K-Means you may find here a fairly simple approach, feel free to send me some feedback if you find it useful or if you need me to add - update things: