jsky
jsky

Reputation: 2217

Convert an images pixel value from rgb to grayscale manually python PIL?

Im trying to use a specific gamma corrected grayscale implementation - Gleam to convert an images pixels to grayscale. How can i do this manually with PIL python?

def tau_gamma_correct(pixel_channel):
    pixel_channel = pixel_channel**(1/2.2)
    return pixel_channel

#@param: rgb
#@result: returns grayscale value
def gleam(rgb):
    #convert rgb tuple to list
    rgblist = list(rgb)
    #gamma correct each rgb channel
    rgblist[0] = tau_gamma_correct(rgblist[0])
    rgblist[1] = tau_gamma_correct(rgblist[1])
    rgblist[2] = tau_gamma_correct(rgblist[2])
    grayscale = 1/3*(rgblist[0] + rgblist[1] + rgblist[2])
    return grayscale

# get a glob list of jpg filenames
files = glob.glob('*.jpg')
for file in files:
    file = open(file)
    filename = file.name
    image = Image.open(file)
    pix = image.load()
    width, height = image.size
    #print(width,height)
    for x in range(0, width):
        for y in range(0, height):
            rgb = pix[x,y]
            #print(rgb)
            # calc new pixel value and set to pixel
            image.mode = 'L'
            pix[x,y] = gleam(rgb)

            image.save(filename + 'gray.gleam'+'.jpg')
   file.close()

SystemError: new style getargs format but argument is not a tuple
It is still expecting the rgb tuple i think.

Upvotes: 3

Views: 2825

Answers (3)

Mark Ransom
Mark Ransom

Reputation: 308166

The problem is that image.mode = 'L' doesn't actually change the type of the image, it just changes the attribute so it's no longer accurate. To change the mode of the image you need to make a new copy with image.convert('L').

Once you have an image in grayscale mode, it won't require a tuple for a pixel value anymore.

Upvotes: 1

jsky
jsky

Reputation: 2217

I found that i could just build another image:

import sys
import os
import glob
import numpy
from PIL import Image

def tau_gamma_correct(pixel_channel):
    pixel_channel = pixel_channel**(1/2.2)
    return pixel_channel

#@param: rgb
#@result: returns grayscale value
def gleam(rgb):
    #convert rgb tuple to list
    rgblist = list(rgb)
    #gamma correct each rgb channel
    rgblist[0] = tau_gamma_correct(rgblist[0])
    print('gleamed red ' + str(rgblist[0]))
    rgblist[1] = tau_gamma_correct(rgblist[1])
    print('gleamed green ' + str(rgblist[1]))
    rgblist[2] = tau_gamma_correct(rgblist[2])
    print('gleamed blue ' + str(rgblist[0]))
    grayscale = (rgblist[0] + rgblist[1] + rgblist[2])/3
    print('grayscale '+ str(grayscale))
    return grayscale

# get a glob list of jpg filenames
files = glob.glob('*.jpg')
for file in files:
    file = open(file)
    filename = file.name
    image = Image.open(file)
    pix = image.load()
    width, height = image.size
    new_image = Image.new('L', image.size)
    #pixelmatrix = [width][height]
    pixelmatrix = numpy.zeros((width, height))
    #print(width,height)
    for x in range(0, width):
        for y in range(0, height):
            rgb = pix[x,y]
            print('current pixel value: '+str(rgb))
            # calc new pixel value and set to pixel
            #print(gleam(rgb))
            gray = gleam(rgb)
            print('changing to pixel value: '+str(gray))
            pixelmatrix[x,y] = gray
            new_image.save(filename + 'gray.gleam'+'.jpg')
    new_image.putdata(pixelmatrix)
    file.close()

Upvotes: 1

ZdaR
ZdaR

Reputation: 22954

Seeing the SystemError: new style getargs format but argument is not a tuple error it seems that you need to return a tuple, which is represented as :

sample_tuple = (1, 2, 3, 4)

So we edit the gleam() function as:

def gleam(rgb):
    #convert rgb tuple to list
    rgblist = list(rgb)
    #gamma correct each rgb channel
    rgblist[0] = tau_gamma_correct(rgblist[0])
    rgblist[1] = tau_gamma_correct(rgblist[1])
    rgblist[2] = tau_gamma_correct(rgblist[2])
    grayscale = 1/3*(rgblist[0] + rgblist[1] + rgblist[2])
    return (grayscale, )

Keep in mind that while returning a single element tuple you need to represent as :

sample_tuple = (1, )

This is due to the fact that (4) == 4 but (4, ) != 4

Upvotes: 0

Related Questions