Converts image input to CIE xy chromaticity coordinates with original pixel colour

picture_to_cie_diagram.py

from PIL import Image
import numpy as np
import matplotlib.pyplot as plt
import time
from colour.plotting import *

def plot_xy_coordinates_with_color(xy_and_rgb_np, output_png_path):
    start_time = time.time()

    xy = xy_and_rgb_np[:, :2]
    RGB = xy_and_rgb_np[:, 2:5]

    if RGB.max() > 1:
        RGB = RGB / 255.0

    plot_chromaticity_diagram_CIE1931(show = False, show_diagram_colours = False)
    
    ax = plt.gca() 
    ax.set_title('CIE 1931 Chromaticity Diagram')
    ax.scatter(xy[:, 0], xy[:, 1], color=RGB, s=0.01, edgecolors=None, linewidths=0)
    
    plt.savefig(output_png_path, format='png', dpi=500)
    plt.close()
    print('Drawing Chromaticity Diagram spent: {:.2f} seconds'.format(time.time() - start_time))

def image_to_cie_xy_and_rgb(image_path):
    start_time = time.time()
    with Image.open(image_path) as img:
        img = img.convert('RGB')
        image_data = np.array(img, dtype=np.float16)

    pixels = image_data.reshape(-1, 3)

    sum_xyz = np.sum(pixels, axis=1, keepdims=True)
    xy_coordinates = np.divide(pixels[:, :2], sum_xyz, out=np.zeros_like(pixels[:, :2], dtype=np.float16), where=sum_xyz!=0)

    xy_and_rgb = np.hstack([xy_coordinates, pixels])


    print('Computing XYZ and RGB spent: {:.2f} seconds'.format(time.time() - start_time))
    
    return xy_and_rgb


if __name__ == "__main__":
    '''
    The Python script converts image RGB data to CIE xy chromaticity coordinates and RGB values
    using PIL and NumPy (function image_to_cie_xy_and_rgb). 
    It then visualizes these coordinates on the CIE 1931 Chromaticity Diagram 
    with Matplotlib and colour.plotting, coloring each point according to 
    its RGB value (function plot_xy_coordinates_with_color). 
    The development focused on optimizing data processing and visualization, 
    utilizing vectorized operations for efficiency, 
    ensuring accurate color representation, and producing visual output.
    '''
    image_path = './test.jpg'
    xy_and_rgb = image_to_cie_xy_and_rgb(image_path)

    output_png_path = 'cie_diagram.png'
    plot_xy_coordinates_with_color(xy_and_rgb, output_png_path)

The following is a set of image samples, for input images:

After processing, you can get:

I had to change the plotting code to put it on the same figure in Jupyter notebook

fig, ax = plt.subplots()
    
  ax.set_title('CIE 1931 Chromaticity Diagram')
  ax.scatter(xy[:, 0], xy[:, 1], color=RGB, s=0.01, edgecolors=None, linewidths=0)
  
  plot_chromaticity_diagram_CIE1931(show=False, show_diagram_colours=False, figure=fig, axes=ax)
  
#     plt.savefig(output_png_path, format='png', dpi=100)
  fig.show()
  plt.close(fig)

@Dawars There are significant logic issues in this code version, causing the graph it generates to inaccurately represent the pixel distribution in the photo within the CIE color space.

My friend @Rachel030219 pointed out these issues and provided the fixed code. You can view the version of the code he modified here. If you want to learn more, please visit his blog (please note that the blog is in Chinese).