palletizer.py

# Author: Teja Gupta, Aditya Gupta
# palletizer.py
# This is a pallete based approach to reduce the size of sprite sheets using kmeans clustering
#
# Anyone using this file is free to use it as long as due credit is given.

# The following imports the necessary libraries make sure you have numpy,matplotlib and sklearn installed.
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import sklearn
from sklearn import cluster
import numpy as np
from PIL import Image

# Opens to file into an Image Object

old_file=input('What is the name of the png you want to palletize (drop the .png): ')
new_file_name=input('Name of the modified image(drop the .png)?: ')
palletized_txt_name=input('Name of the palletized labels(drop the file extension): ')
pallete_file=input('Name of the pallete(drop the file extension)?: ')
color_number=int(input('How many colors do you want?: '))

im=Image.open(old_file+'.png')
im.convert('RGBA')

# Converts the image to a list of pixels
list_=[]
for x in range(im.size[0]):
	for y in range(im.size[1]):
		list_.append(list(im.getpixel((x,y))[:3]))

# Performs KMeans Clustering
print('Doing KMeans Clustering...')
kmean=cluster.KMeans(n_clusters=color_number).fit(list_)
print('...Finished KMeans Clustering')

# Make a new image and file to save the picture
new_image = Image.new('RGB', im.size, color='white')

# Make a new file to save the indices into the palette
new_file= open(palletized_txt_name+'.txt','w')

# Gets the labels and cluster centers (which become the pallete)
labels=np.array(kmean.labels_)
centers=np.array(kmean.cluster_centers_,dtype='uint8');

# Makes the new image and file based on the pallete
for x in range(im.size[0]):
	for y in range(im.size[1]):
		new_image.putpixel((x,y),tuple(centers[labels[x*im.size[1]+y]]))
		new_file.write('%d\n'%(labels[x*im.size[1]+y]))
new_file.close()
new_image.save(new_file_name+'.png')

# Saves the cluster centers as the pallete
new_pallete=open(pallete_file+'.txt','w')
for i in range(len(centers)):
	new_pallete.write(('0x%02x%02x%02x\n')%(centers[i][0],centers[i][1],centers[i][2]))

new_pallete.close()