kmeans

Ik = kmeans(I::GMTimage, k=5; seeds=nothing, maxiter=100, tol=1e-7, V=false) -> GMTimage

Compute a k-means clustering on an RGB image I. It produces a fixed number of clusters, each associated with a center, and each RGB color is assigned to a cluster with the nearest center.

I: The input \(GMTimage\) object.
k: The number of clusters when using unsupervised classification.
seeds: For supervised classifications this is Mx3 UInt8 matrix with the colors of the cluster centers. The algorithm than aggregates all colors in the image around these M seed colors. Attention, if provided, this option resets k.
maxiter: Maximum number of iterations that the algorithm may run till reach a solution.
tol: Alternatively, sets the minimal allowed change of the objective during convergence. The iteration process stops when one of the two conditions is met first.
V: Print some info at the end of the iterative loop (number of iterations, time spent).

kmeans(X::Union{GMTdataset, Matrix{<:Real}}, k=3; seeds=nothing, maxiter=100, tol=1e-7,
       raw::Bool=false, V=false) -> Vector{GMTdataset} | idx, centers, counts

This method accepts a M-by-d matrix or a \(GMTdataset\) where columns represent the data points and rows the d-dimensional data point.

raw: A Boolean that if false makes the return data be a vector of \(GMTdatset\), one for each cluster found in input data. If raw=true, we return: idx, centers, counts, where
- idx: A vector of ints with the assignments of each data points (by position in the idx vector) to clusters.
- centers: A k-by-d matrix with the centers of each cluster.
- counts: A matrix of integers with the cluster number in first column, and number of elements in that cluster in second column.

Example

    D = gmtread(TESTSDIR * "iris.dat");
    Dk = kmeans(D, k=3)     # Unsupervised segment data into 3 clusters.

Methods

# 4 methods for generic function "kmeans" from GMT:
 [1] kmeans(X::Union{GMTdataset, Matrix{<:Real}}, k; seeds, maxiter, tol, raw, V)
 [2] kmeans(X::Union{GMTdataset, Matrix{<:Real}}; ...)
 [3] kmeans(I::GMTimage, k; seeds, maxiter, tol, V)
 [4] kmeans(I::GMTimage; ...)

Example

Methods

See Also