imgIsoData {biOps}R Documentation

Image clustering

Description

This function performs an unsupervised classification through the k-means algorithm. It is an enhanced implementation, that avoid some comparisons based on kept information about distances and centroids of previous iterations.

Usage

imgIsoData (imgdata, k, min_dist=1, min_elems=1, split_sd=0.1, iter_start=5, max_merge=2, max_iter=10)

Arguments

imgdata The image
k Number of clusters
min_dist Minimum distance between cluster centroids
min_elems Minimum elements per cluster
split_sd Standard deviation threshold for splitting operation
iter_start Maximum number of forgy iterations
max_merge Maximum of merge operations per iteration
max_iter Maximum number of iterations

Value

return an imagedata object, the result of the classification

See Also

imgKMeans imgEKMeans imgKDKMeans

Examples

        ## Not run: 
                x <- readJpeg(system.file("samples", "violet.jpg", package="biOps"))
                y <- imgIsoData(x, 4)
        
## End(Not run)

[Package biOps version 0.2.1 Index]