CODEWITHSUNDEEP
rimageimage-processingcolorsrgb
dag
dag

Reputation: 59

Image colors composition using R

I'm trying to obtain the percentage of each color which compose a given picture. I get the RGB matrix using this code:

library(jpeg)
img <- readJPEG("C:/Users/Pictures/img.jpg")
dim(img)
#145 371   3
img<-img*255
#this operation is necessary to obtain an RBG scale

At this step I'm not sure which is the right way to go ahead. Anyway, I would like to obtain someting like this:

Count    RGB vector

200614   (255,255,255) 

4758     (253,253,218) 

4312     (250,250,229) 

1821     (235,237,242) 

1776     (212,214,226)

...

and then I can calculate the percentage of each color. Lastly I'll try to associate a label to each RGB vector.

Anyone can help me?

Upvotes: 1

Views: 1751

Answers (1)

aoles
aoles

Reputation: 1545

You can easily count colors with table after converting the pixel array to #rrggbb values with as.raster (note that you don't need to scale the values by multiplying them by 255). Individual color components from hex color strings can be obtained by col2rgb.

library(jpeg)
img <- readJPEG("C:/Users/Pictures/img.jpg")

# Convert the 3D array to a matrix of #rrggbb values
img <- as.raster(img)

# Create a count table
tab <- table(img)

# Convert to a data.frame
tab <- data.frame(Color = names(tab), Count = as.integer(tab))

# Extract red/green/blue values
RGB <- t(col2rgb(tab$Color))
tab <- cbind(tab, RGB)

# Preview
head(tab)

Upvotes: 9

Related Questions