Description of K-Means Point Reduction Algorithm
- Date 20/07/2017.
- Author: Guillaume Touya
- Contact {firstname.lastname}@ign.fr.
Description of the algorithm
The algorithm computes a K-Mean clustering of the points using the Euclidean distance between points, and then replaces each cluster by a single point.
More details from Wikipedia on K-Means clustering.
| Parameter name | Description | Type | Default value |
|---|---|---|---|
| k | the number of clusters to build (i.e. the number of points after reduction | integer | |
| shrinkRatio | if k is set to 0, the algorithm rather uses this ratio to compute how many points are kept (0.5 means half the points are kept) | double (between 0 and 1) |
Examples of generalization
When to use the algorithm?
The algorithm is dedicated to features represented by point clouds.