提取图片主要颜色脚本

extra_color.py

import matplotlib.image as img
import pandas as pd
from scipy.cluster.vq import whiten, kmeans

target_image = img.imread("path/to/target_image")
r = []
g = []
b = []
for row in target_image:
    for temp_r, temp_g, temp_b in row:
        r.append(temp_r)
        g.append(temp_g)
        b.append(temp_b)

image_df = pd.DataFrame({'red': r,
                         'green': g,
                         'blue': b})

# scale image information
image_df['scaled_color_red'] = whiten(image_df['red'])
image_df['scaled_color_blue'] = whiten(image_df['blue'])
image_df['scaled_color_green'] = whiten(image_df['green'])

cluster_centers, _ = kmeans(image_df[['scaled_color_red',
                                      'scaled_color_blue',
                                      'scaled_color_green']], 3)

dominant_colors = []

# Get standard deviations of each color
red_std, green_std, blue_std = image_df[['red',
                                         'green',
                                         'blue']].std()

for cluster_center in cluster_centers:
    red_scaled, green_scaled, blue_scaled = cluster_center

    # Convert each standardized value to scaled value
    dominant_colors.append((
        red_scaled * red_std,
        green_scaled * green_std,
        blue_scaled * blue_std
    ))

print(dominant_colors)

使用的是 scipy 的 kmeans 函数，对于比较大的图片可能要算上蛮久

实际上可以先对图片大小进行压缩，例如压缩到 200x200的尺寸，颜色信息相差也不会很大