image_clustering.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf \n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics import silhouette_score\n",
    "import cv2\n",
    "import os, glob, shutil"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Run this cell if you have issues with intitializing cuDNN\n",
    "gpus = tf.config.experimental.list_physical_devices('GPU')\n",
    "if gpus:\n",
    "  try:\n",
    "    # Currently, memory growth needs to be the same across GPUs\n",
    "    for gpu in gpus:\n",
    "      tf.config.experimental.set_memory_growth(gpu, True)\n",
    "    logical_gpus = tf.config.experimental.list_logical_devices('GPU')\n",
    "    print(len(gpus), \"Physical GPUs,\", len(logical_gpus), \"Logical GPUs\")\n",
    "  except RuntimeError as e:\n",
    "    # Memory growth must be set before GPUs have been initialized\n",
    "    print(e)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_dir = 'pets'\n",
    "glob_dir = input_dir + '/*.jpg'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "images = [cv2.resize(cv2.imread(file), (224, 224)) for file in glob.glob(glob_dir)]\n",
    "paths = [file for file in glob.glob(glob_dir)]\n",
    "images = np.array(np.float32(images).reshape(len(images), -1)/255)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 1s 671ms/step\n"
     ]
    }
   ],
   "source": [
    "model = tf.keras.applications.MobileNetV2(include_top=False, weights='imagenet', input_shape=(224, 224, 3))\n",
    "predictions = model.predict(images.reshape(-1, 224, 224, 3))\n",
    "pred_images = predictions.reshape(images.shape[0], -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n",
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n"
     ]
    }
   ],
   "source": [
    "sil = []\n",
    "kl = []\n",
    "kmax = 10\n",
    "\n",
    "for k in range(2, kmax+1):\n",
    "  kmeans2 = KMeans(n_clusters = k).fit(pred_images)\n",
    "  labels = kmeans2.labels_\n",
    "  sil.append(silhouette_score(pred_images, labels, metric = 'euclidean'))\n",
    "  kl.append(k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(kl, sil)\n",
    "plt.ylabel('Silhoutte Score')\n",
    "plt.ylabel('K')\n",
    "plt.show()\n",
    "\n",
    "# You can decide the optimal value of K. For the pets example, its 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Python310\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n"
     ]
    }
   ],
   "source": [
    "k = 2\n",
    "kmodel = KMeans(n_clusters=k, random_state=728)\n",
    "kmodel.fit(pred_images)\n",
    "kpredictions = kmodel.predict(pred_images)\n",
    "shutil.rmtree('output')\n",
    "for i in range(k):\n",
    "\tos.makedirs(\"output\\cluster\" + str(i))\n",
    "for i in range(len(paths)):\n",
    "\tshutil.copy2(paths[i], \"output\\cluster\"+str(kpredictions[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}