Mural Color Analysis PB

Color Analysis

This document visualizes K-means clustering on the colors of a particular image design to determine the most important colors needed.

Setup The Function

get_plots <- function(img_path, kk) {
  
  mural <- readJPEG(img_path)
  # reshape image into a data frame
  df = data.frame(
    red = matrix(mural[,,1], ncol=1),
    green = matrix(mural[,,2], ncol=1),
    blue = matrix(mural[,,3], ncol=1)
  )
  
  ### compute the k-means clustering
  K = kmeans(df,kk)
  df$label = K$cluster
  
  ### Replace the color of each pixel in the image with the mean 
  ### R,G, and B values of the cluster in which the pixel resides:
  
  # get the coloring
  colors = data.frame(
    label = 1:nrow(K$centers), 
    R = K$centers[,"red"],
    G = K$centers[,"green"],
    B = K$centers[,"blue"]
  )
  
  # merge color codes on to df
  # IMPORTANT: we must maintain the original order of the df after the merge!
  df$order = 1:nrow(df)
  df = merge(df, colors)
  df = df[order(df$order),]
  df$order = NULL

  #Get Hex Colors and Percentage of Each
  kct <- df %>%
    mutate(hex = rgb(R,G,B)) %>%
    group_by(hex) %>%
    summarise(ct = n()) %>%
    mutate(pct = scales::percent(ct/sum(ct)))
  
  
  # get mean color channel values for each row of the df.
  R = matrix(df$R, nrow=dim(mural)[1])
  G = matrix(df$G, nrow=dim(mural)[1])
  B = matrix(df$B, nrow=dim(mural)[1])
  
  # reconstitute the segmented image in the same shape as the input image
  mural.segmented = array(dim=dim(mural))
  mural.segmented[,,1] = R
  mural.segmented[,,2] = G
  mural.segmented[,,3] = B
  
  img_S = rasterGrob(mural.segmented)
  
  
  return(list(#original_plot=g1,segmented_plot = g2,
    segmented_img=img_S,
    Clrs = kct,
    TWSS = K$tot.withinss
    ))
}

##Execute Run the defined Function to get tables and Images

pt <- proc.time()
w <- as.numeric()
for (i in 1:20) {
  g <- get_plots("/Users/newuser/Downloads/mur2.jpeg",i)
  grid.arrange(g[[1]], nrow=1)
  ggsave(paste("seg_Outp_img4_k",i,".png",sep=""), g[[1]])
  cat("\n") 
  cat("## Num Colors: ", i, "\n")
  print(
   tagList(
    datatable(g[[2]])
   )
  )
  
  cat("\n")
  
  
  w[i] <- g[[3]]
}

## Saving 7 x 5 in image

## Num Colors: 1

## Saving 7 x 5 in image

## Num Colors: 2

## Saving 7 x 5 in image

## Num Colors: 3

## Saving 7 x 5 in image

## Num Colors: 4

## Saving 7 x 5 in image

## Num Colors: 5

## Saving 7 x 5 in image

## Num Colors: 6

## Saving 7 x 5 in image

## Num Colors: 7

## Saving 7 x 5 in image

## Num Colors: 8

## Saving 7 x 5 in image

Num Colors: 9

## Warning: did not converge in 10 iterations

## Saving 7 x 5 in image

Num Colors: 10

## Warning: did not converge in 10 iterations

## Saving 7 x 5 in image

## Num Colors: 11

## Saving 7 x 5 in image

## Num Colors: 12

## Saving 7 x 5 in image

## Num Colors: 13

## Saving 7 x 5 in image

## Num Colors: 14

## Saving 7 x 5 in image

## Num Colors: 15

## Saving 7 x 5 in image

## Num Colors: 16

## Saving 7 x 5 in image

Num Colors: 17

## Warning: did not converge in 10 iterations

## Saving 7 x 5 in image

Num Colors: 18

## Warning: did not converge in 10 iterations

## Saving 7 x 5 in image

## Num Colors: 19

## Saving 7 x 5 in image

## Num Colors: 20

How Well Do The Clusters Fit?

wdf <- data.frame(k=1:10, wss = w)

gg <- ggplot(wdf, aes(x=k, y=wss)) +
  geom_line() +
  geom_point()
gg