{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook shows primary color analysis of color image using K-Means algorithm.\n",
    "The output are N primary colors and their corresponding percentage."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "import matplotlib.pyplot as plt\n",
    "import cv2\n",
    "import numpy as np\n",
    "import requests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def centroid_histogram(clt):\n",
    "    # grab the number of different clusters and create a histogram\n",
    "    # based on the number of pixels assigned to each cluster\n",
    "    numLabels = np.arange(0, len(np.unique(clt.labels_)) + 1)\n",
    "    (hist, _) = np.histogram(clt.labels_, bins=numLabels)\n",
    "\n",
    "    # normalize the histogram, such that it sums to one\n",
    "    hist = hist.astype(\"float\")\n",
    "    hist /= hist.sum()\n",
    "\n",
    "    # return the histogram\n",
    "    return hist\n",
    "\n",
    "\n",
    "def plot_colors(hist, centroids):\n",
    "    # initialize the bar chart representing the relative frequency\n",
    "    # of each of the colors\n",
    "    bar = np.zeros((50, 300, 3), dtype=\"uint8\")\n",
    "    startX = 0\n",
    "    # loop over the percentage of each cluster and the color of\n",
    "    # each cluster\n",
    "    for (percent, color) in zip(hist, centroids):\n",
    "        # plot the relative percentage of each cluster\n",
    "        endX = startX + (percent * 300)\n",
    "        cv2.rectangle(\n",
    "            bar, (int(startX), 0), (int(endX), 50), color.astype(\"uint8\").tolist(), -1\n",
    "        )\n",
    "        startX = endX\n",
    "\n",
    "    # return the bar chart\n",
    "    return bar"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load the image and convert it from BGR to RGB so that\n",
    "# we can dispaly it with matplotlib\n",
    "# image_path = './data/blue.jpg'\n",
    "# image = cv2.imread(image_path)\n",
    "\n",
    "file = requests.get(\n",
    "    \"https://heibox.uni-heidelberg.de/thumbnail/537e6da0a8b44069bc96/1024/images/100361_asm.png\"\n",
    ")\n",
    "image = cv2.imdecode(np.fromstring(file.content, np.uint8), 1)\n",
    "\n",
    "# BGR-->RGB cv to matplotlib show\n",
    "image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n",
    "\n",
    "# show our image\n",
    "plt.figure()\n",
    "plt.axis(\"off\")\n",
    "plt.imshow(image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reshape the image to be a list of pixels\n",
    "image = image.reshape((image.shape[0] * image.shape[1], 3))\n",
    "\n",
    "# cluster the pixel intensities\n",
    "clt = KMeans(n_clusters=8)\n",
    "clt.fit(image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# build a histogram of clusters and then create a figure\n",
    "# representing the number of pixels labeled to each color\n",
    "hist = centroid_histogram(clt)\n",
    "bar = plot_colors(hist, clt.cluster_centers_)\n",
    "\n",
    "# show our color bart\n",
    "plt.figure()\n",
    "plt.axis(\"off\")\n",
    "plt.imshow(bar)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for (percent, color) in zip(hist, clt.cluster_centers_):\n",
    "    print(\"color:\", color, \" percentage:\", percent)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}