Mean Shift
Reference work entry
DOI: https://doi.org/10.1007/978-0-387-30164-8_527
Mean shift (Comaniciu & Meer, 2002) is a nonparametric algorithm for partitional clustering which does not require specifying the number of clusters, and can form any shape of clusters.
Given
n data points
x
i,
i = 1,
…,
n, in the
d-dimensional space
R
d, the multivariate kernel density estimator obtained with kernel
K(
x) and window radius
h is given by
$$f(x) = \frac{1} {n{h}^{d}}{ \sum \limits _{i=1}^{n}}K\left (\frac{x - {x}_{i}} {h} \right ).$$
This is a preview of subscription content, log in to check access.
Recommended Reading
- Comaniciu, D., & Meer, P. (2002). Mean shift: A robust approach toward feature space analysis. IEEE Transactions of the Pattern Analysis and Machine Intelligence, 24(5), 603–619.Google Scholar
Copyright information
© Springer Science+Business Media, LLC 2011