image_kmeans: Obtain the Matrix of the Segmented Image using K-means...

In rwang14/implant: A High-throughput Phenotyping Pipeline for Image Processing and Functional Growth Curve Analysis

Description

This function is to obtain a segmented plant image using K-means Clustering Method, together with the means and standard deviations of the pixel intensities for different classes.

Usage

image_kmeans(Y, k)

Arguments

Y	an input matrix. For plant segmentation, we recommend to use the relative green intensity of the image.
k	a positive integer, which refers to the cluster number. In default, k = 2 to seperate the plant from the background.

Value

X	a matrix of the class label of the segmented image.
mu	a k by 1 matrix. Each row represents the sample mean of each cluster.
sigma	a k by 1 matrix. Each row represents the sample standard deviation of each cluster.

Examples

library(png)
orig = readPNG(system.file("extdata", "reduced.png", package = "implant", mustWork = TRUE))
#Define the response as relative green.
Y = orig[ , , 2]/(orig[ , , 1]+orig[ , , 2]+orig[ , , 3])
#Take the initial label of EM algorithm using K-means
X = image_kmeans(Y, k = 2)$X
#Obtain the image produced by kmeans clustering
output = matrix(as.numeric(X), nrow = nrow(X), ncol = ncol(X)) - 1
writePNG(output,"~/kmeans.png")

rwang14/implant documentation built on Sept. 6, 2020, 3:21 a.m.