face_recognition.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "# Faces recognition example using eigenfaces and SVMs\n",
    "\n",
    "\n",
    "The dataset used in this example is a preprocessed excerpt of the\n",
    "\"Labeled Faces in the Wild\", aka LFW_:\n",
    "\n",
    "  http://vis-www.cs.umass.edu/lfw/lfw-funneled.tgz (233MB)\n",
    "\n",
    ".. _LFW: http://vis-www.cs.umass.edu/lfw/\n",
    "\n",
    "Expected results for the top 5 most represented people in the dataset:\n",
    "\n",
    "================== ============ ======= ========== =======\n",
    "                   precision    recall  f1-score   support\n",
    "================== ============ ======= ========== =======\n",
    "     Ariel Sharon       0.67      0.92      0.77        13\n",
    "     Colin Powell       0.75      0.78      0.76        60\n",
    "  Donald Rumsfeld       0.78      0.67      0.72        27\n",
    "    George W Bush       0.86      0.86      0.86       146\n",
    "Gerhard Schroeder       0.76      0.76      0.76        25\n",
    "      Hugo Chavez       0.67      0.67      0.67        15\n",
    "       Tony Blair       0.81      0.69      0.75        36\n",
    "\n",
    "      avg / total       0.80      0.80      0.80       322\n",
    "================== ============ ======= ========== =======\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "ImportError",
     "evalue": "No module named model_selection",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mImportError\u001b[0m                               Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-1-f1e519288a81>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      5\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mmatplotlib\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpyplot\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      6\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 7\u001b[1;33m \u001b[1;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmodel_selection\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mtrain_test_split\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      8\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmodel_selection\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mGridSearchCV\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      9\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdatasets\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mfetch_lfw_people\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mImportError\u001b[0m: No module named model_selection"
     ]
    }
   ],
   "source": [
    "from __future__ import print_function\n",
    "\n",
    "from time import time\n",
    "import logging\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.datasets import fetch_lfw_people\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.svm import SVC\n",
    "\n",
    "\n",
    "print(__doc__)\n",
    "\n",
    "# Display progress logs on stdout\n",
    "logging.basicConfig(level=logging.INFO, format='%(asctime)s %(message)s')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Download the data, if not already on disk and load it as numpy arrays\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "lfw_people = fetch_lfw_people(min_faces_per_person=70, resize=0.4)\n",
    "\n",
    "# introspect the images arrays to find the shapes (for plotting)\n",
    "n_samples, h, w = lfw_people.images.shape\n",
    "\n",
    "# for machine learning we use the 2 data directly (as relative pixel\n",
    "# positions info is ignored by this model)\n",
    "X = lfw_people.data\n",
    "n_features = X.shape[1]\n",
    "\n",
    "# the label to predict is the id of the person\n",
    "y = lfw_people.target\n",
    "target_names = lfw_people.target_names\n",
    "n_classes = target_names.shape[0]\n",
    "\n",
    "print(\"Total dataset size:\")\n",
    "print(\"n_samples: %d\" % n_samples)\n",
    "print(\"n_features: %d\" % n_features)\n",
    "print(\"n_classes: %d\" % n_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split into a training set and a test set using a stratified k fold\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# split into a training and testing set\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.25, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compute a PCA (eigenfaces) on the face dataset (treated as unlabeled\n",
    "dataset): unsupervised feature extraction / dimensionality reduction\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "n_components = 150\n",
    "\n",
    "print(\"Extracting the top %d eigenfaces from %d faces\"\n",
    "      % (n_components, X_train.shape[0]))\n",
    "t0 = time()\n",
    "pca = PCA(n_components=n_components, svd_solver='randomized',\n",
    "          whiten=True).fit(X_train)\n",
    "print(\"done in %0.3fs\" % (time() - t0))\n",
    "\n",
    "eigenfaces = pca.components_.reshape((n_components, h, w))\n",
    "\n",
    "print(\"Projecting the input data on the eigenfaces orthonormal basis\")\n",
    "t0 = time()\n",
    "X_train_pca = pca.transform(X_train)\n",
    "X_test_pca = pca.transform(X_test)\n",
    "print(\"done in %0.3fs\" % (time() - t0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train a SVM classification model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "print(\"Fitting the classifier to the training set\")\n",
    "t0 = time()\n",
    "param_grid = {'C': [1e3, 5e3, 1e4, 5e4, 1e5],\n",
    "              'gamma': [0.0001, 0.0005, 0.001, 0.005, 0.01, 0.1], }\n",
    "clf = GridSearchCV(SVC(kernel='rbf', class_weight='balanced'), param_grid)\n",
    "clf = clf.fit(X_train_pca, y_train)\n",
    "print(\"done in %0.3fs\" % (time() - t0))\n",
    "print(\"Best estimator found by grid search:\")\n",
    "print(clf.best_estimator_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quantitative evaluation of the model quality on the test set\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "print(\"Predicting people's names on the test set\")\n",
    "t0 = time()\n",
    "y_pred = clf.predict(X_test_pca)\n",
    "print(\"done in %0.3fs\" % (time() - t0))\n",
    "\n",
    "print(classification_report(y_test, y_pred, target_names=target_names))\n",
    "print(confusion_matrix(y_test, y_pred, labels=range(n_classes)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Qualitative evaluation of the predictions using matplotlib\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def plot_gallery(images, titles, h, w, n_row=3, n_col=4):\n",
    "    \"\"\"Helper function to plot a gallery of portraits\"\"\"\n",
    "    plt.figure(figsize=(1.8 * n_col, 2.4 * n_row))\n",
    "    plt.subplots_adjust(bottom=0, left=.01, right=.99, top=.90, hspace=.35)\n",
    "    for i in range(n_row * n_col):\n",
    "        plt.subplot(n_row, n_col, i + 1)\n",
    "        plt.imshow(images[i].reshape((h, w)), cmap=plt.cm.gray)\n",
    "        plt.title(titles[i], size=12)\n",
    "        plt.xticks(())\n",
    "        plt.yticks(())\n",
    "\n",
    "\n",
    "# plot the result of the prediction on a portion of the test set\n",
    "\n",
    "def title(y_pred, y_test, target_names, i):\n",
    "    pred_name = target_names[y_pred[i]].rsplit(' ', 1)[-1]\n",
    "    true_name = target_names[y_test[i]].rsplit(' ', 1)[-1]\n",
    "    return 'predicted: %s\\ntrue:      %s' % (pred_name, true_name)\n",
    "\n",
    "prediction_titles = [title(y_pred, y_test, target_names, i)\n",
    "                     for i in range(y_pred.shape[0])]\n",
    "\n",
    "plot_gallery(X_test, prediction_titles, h, w)\n",
    "\n",
    "# plot the gallery of the most significative eigenfaces\n",
    "\n",
    "eigenface_titles = [\"eigenface %d\" % i for i in range(eigenfaces.shape[0])]\n",
    "plot_gallery(eigenfaces, eigenface_titles, h, w)\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}