Traffic_Sign_Classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "# Self-Driving Car Engineer Nanodegree\n",
    "\n",
    "## Deep Learning\n",
    "\n",
    "## Project: Build a Traffic Sign Recognition Classifier\n",
    "\n",
    "In this notebook, a template is provided for you to implement your functionality in stages, which is required to successfully complete this project. If additional code is required that cannot be included in the notebook, be sure that the Python code is successfully imported and included in your submission if necessary. \n",
    "\n",
    "> **Note**: Once you have completed all of the code implementations, you need to finalize your work by exporting the iPython Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to  \\n\",\n",
    "    \"**File -> Download as -> HTML (.html)**. Include the finished document along with this notebook as your submission. \n",
    "\n",
    "In addition to implementing code, there is a writeup to complete. The writeup should be completed in a separate file, which can be either a markdown file or a pdf document. There is a [write up template](https://github.com/udacity/CarND-Traffic-Sign-Classifier-Project/blob/master/writeup_template.md) that can be used to guide the writing process. Completing the code template and writeup template will cover all of the [rubric points](https://review.udacity.com/#!/rubrics/481/view) for this project.\n",
    "\n",
    "The [rubric](https://review.udacity.com/#!/rubrics/481/view) contains \"Stand Out Suggestions\" for enhancing the project beyond the minimum requirements. The stand out suggestions are optional. If you decide to pursue the \"stand out suggestions\", you can include the code in this Ipython notebook and also discuss the results in the writeup file.\n",
    "\n",
    "\n",
    ">**Note:** Code and Markdown cells can be executed using the **Shift + Enter** keyboard shortcut. In addition, Markdown cells can be edited by typically double-clicking the cell to enter edit mode."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "---\n",
    "## Step 0: Load The Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Load pickled data\n",
    "import pickle\n",
    "\n",
    "## bringing in the train, validation, and test data\n",
    "# training_file = '/Users/jwen/Python/Projects/sdcnd_traffic_sign/traffic-signs-data/train.p'\n",
    "# validation_file= '/Users/jwen/Python/Projects/sdcnd_traffic_sign/traffic-signs-data/valid.p'\n",
    "# testing_file = '/Users/jwen/Python/Projects/sdcnd_traffic_sign/traffic-signs-data/test.p'\n",
    "\n",
    "training_file = '/home/carnd/sdcnd_traffic_sign/traffic-signs-data/train.p'\n",
    "validation_file= '/home/carnd/sdcnd_traffic_sign/traffic-signs-data/valid.p'\n",
    "testing_file = '/home/carnd/sdcnd_traffic_sign/traffic-signs-data/test.p'\n",
    "\n",
    "with open(training_file, mode='rb') as f:\n",
    "    train = pickle.load(f)\n",
    "with open(validation_file, mode='rb') as f:\n",
    "    valid = pickle.load(f)\n",
    "with open(testing_file, mode='rb') as f:\n",
    "    test = pickle.load(f)\n",
    "    \n",
    "X_train, y_train = train['features'], train['labels']\n",
    "X_valid, y_valid = valid['features'], valid['labels']\n",
    "X_test, y_test = test['features'], test['labels']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "---\n",
    "\n",
    "## Step 1: Dataset Summary & Exploration\n",
    "\n",
    "The pickled data is a dictionary with 4 key/value pairs:\n",
    "\n",
    "- `'features'` is a 4D array containing raw pixel data of the traffic sign images, (num examples, width, height, channels).\n",
    "- `'labels'` is a 1D array containing the label/class id of the traffic sign. The file `signnames.csv` contains id -> name mappings for each id.\n",
    "- `'sizes'` is a list containing tuples, (width, height) representing the original width and height the image.\n",
    "- `'coords'` is a list containing tuples, (x1, y1, x2, y2) representing coordinates of a bounding box around the sign in the image. **THESE COORDINATES ASSUME THE ORIGINAL IMAGE. THE PICKLED DATA CONTAINS RESIZED VERSIONS (32 by 32) OF THESE IMAGES**\n",
    "\n",
    "Complete the basic data summary below. Use python, numpy and/or pandas methods to calculate the data summary rather than hard coding the results. For example, the [pandas shape method](http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.shape.html) might be useful for calculating some of the summary results. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Provide a Basic Summary of the Data Set Using Python, Numpy and/or Pandas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training examples = 34799\n",
      "Number of validation examples = 4410\n",
      "Number of testing examples = 12630\n",
      "Image data shape = (32, 32)\n",
      "Number of classes = 43\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "# Number of training examples\n",
    "n_train = train['features'].shape[0]\n",
    "\n",
    "# Number of testing examples.\n",
    "n_test = test['features'].shape[0]\n",
    "\n",
    "# What's the shape of an traffic sign image?\n",
    "image_shape = train['features'].shape[1:3]\n",
    "\n",
    "# How many unique classes/labels there are in the dataset.\n",
    "n_classes = len(np.unique(train['labels']))\n",
    "\n",
    "print(\"Number of training examples =\", n_train)\n",
    "print(\"Number of validation examples =\", valid['features'].shape[0])\n",
    "print(\"Number of testing examples =\", n_test)\n",
    "print(\"Image data shape =\", image_shape)\n",
    "print(\"Number of classes =\", n_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Include an exploratory visualization of the dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Visualize the German Traffic Signs Dataset using the pickled file(s). This is open ended, suggestions include: plotting traffic sign images, plotting the count of each sign, etc.\n",
    "\n",
    "The [Matplotlib](http://matplotlib.org/) [examples](http://matplotlib.org/examples/index.html) and [gallery](http://matplotlib.org/gallery.html) pages are a great resource for doing visualizations in Python.\n",
    "\n",
    "**NOTE:** It's recommended you start with something simple first. If you wish to do more, come back to it after you've completed the rest of the sections."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "### Data exploration visualization code goes here.\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import random\n",
    "%matplotlib inline\n",
    "\n",
    "sign_names = pd.read_csv('/Users/jwen/Python/Projects/sdcnd_traffic_sign/signnames.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 608,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x14f661128>"
      ]
     },
     "execution_count": 608,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATAAAAE8CAYAAABHOe2IAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x15c622ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## plotting a sample traffic sign\n",
    "plt.figure(figsize=(5,5))\n",
    "plt.grid(False)\n",
    "plt.title(str(sign_names[sign_names[\"ClassId\"] == y_train[index]].iloc[0,1]) + \" - Class \" + str(class_num))\n",
    "plt.tick_params(axis='x', colors=(0,0,0,0))\n",
    "plt.tick_params(axis='y', colors=(0,0,0,0))\n",
    "plt.imshow(X_train[1099])\n",
    "\n",
    "## save the image\n",
    "# plt.savefig('/Users/jwen/Python/Projects/sdcnd_traffic_sign/media/class0_class4_2.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 609,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x15b3a9a58>"
      ]
     },
     "execution_count": 609,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAETCAYAAAAs4pGmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJBUZEJKkIgNCklRkQEiSiv4/NASDwPw0qB4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x14fd5c4a8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## histogram of the different sign types\n",
    "plt.hist(train['labels'])\n",
    "plt.xlabel('Sign Index')\n",
    "plt.ylabel('Count')\n",
    "plt.title('Histogram of Sign Types')\n",
    "\n",
    "# fig = plt.figure(figsize=(10,6))\n",
    "# ax = fig.add_subplot(111)\n",
    "# plt.hist(train['labels'])\n",
    "# plt.xlabel('Sign Index')\n",
    "# plt.ylabel('Count')\n",
    "# plt.title('Histogram of Sign Types')\n",
    "# plt.savefig('/Users/jwen/Python/Projects/sdcnd_traffic_sign/media/histogram.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "----\n",
    "\n",
    "## Step 2: Design and Test a Model Architecture\n",
    "\n",
    "Design and implement a deep learning model that learns to recognize traffic signs. Train and test your model on the [German Traffic Sign Dataset](http://benchmark.ini.rub.de/?section=gtsrb&subsection=dataset).\n",
    "\n",
    "The LeNet-5 implementation shown in the [classroom](https://classroom.udacity.com/nanodegrees/nd013/parts/fbf77062-5703-404e-b60c-95b78b2f3f9e/modules/6df7ae49-c61c-4bb2-a23e-6527e69209ec/lessons/601ae704-1035-4287-8b11-e2c2716217ad/concepts/d4aca031-508f-4e0b-b493-e7b706120f81) at the end of the CNN lesson is a solid starting point. You'll have to change the number of classes and possibly the preprocessing, but aside from that it's plug and play! \n",
    "\n",
    "With the LeNet-5 solution from the lecture, you should expect a validation set accuracy of about 0.89. To meet specifications, the validation set accuracy will need to be at least 0.93. It is possible to get an even higher accuracy, but 0.93 is the minimum for a successful project submission. \n",
    "\n",
    "There are various aspects to consider when thinking about this problem:\n",
    "\n",
    "- Neural network architecture (is the network over or underfitting?)\n",
    "- Play around preprocessing techniques (normalization, rgb to grayscale, etc)\n",
    "- Number of examples per label (some have more than others).\n",
    "- Generate fake data.\n",
    "\n",
    "Here is an example of a [published baseline model on this problem](http://yann.lecun.com/exdb/publis/pdf/sermanet-ijcnn-11.pdf). It's not required to be familiar with the approach used in the paper but, it's good practice to try to read papers like these."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Pre-process the Data Set (normalization, grayscale, etc.)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Use the code cell (or multiple code cells, if necessary) to implement the first step of your project."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "## Shuffling the data\n",
    "from sklearn.utils import shuffle\n",
    "\n",
    "X_train, y_train = shuffle(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## preprocessing steps\n",
    "import cv2\n",
    "import random\n",
    "from scipy import ndimage\n",
    "from tqdm import tqdm\n",
    "import pandas as pd\n",
    "\n",
    "# Grayscale because the color dimensions add complexity and might not provide much information\n",
    "def grayscale(img):\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
    "    return img\n",
    "\n",
    "# Normalizes the data to have a mean of 0 and unit variance\n",
    "def normalize(img):\n",
    "    return (img - np.mean(img))/np.std(img)\n",
    "#     return (img - 128)/128\n",
    "#     return img / 255 * 0.8 + 0.1\n",
    "    \n",
    "# Iterates through and preprocesses the images\n",
    "def preprocess(img_set):\n",
    "    new_img_set = []\n",
    "    for img in img_set:\n",
    "        gray_img = grayscale(img)\n",
    "        normalize_img = normalize(gray_img)\n",
    "        new_img_set.append(normalize_img)\n",
    "        \n",
    "    return np.array(new_img_set)\n",
    "\n",
    "## create function for rotating images\n",
    "def rotate_image(img):\n",
    "    angle_list = [-15, -10, -5, 5, 10, 15]\n",
    "    return ndimage.interpolation.rotate(img, random.choice(angle_list), mode='nearest')\n",
    "\n",
    "## zooms into the image (FROM STACKOVERFLOW: \n",
    "## http://stackoverflow.com/questions/37119071/scipy-rotate-and-zoom-an-image-without-changing-its-dimensions/37121993#37121993)\n",
    "def clipped_zoom(img, **kwargs):\n",
    "\n",
    "    zoom_factor = random.choice([1.1, 1.2])\n",
    "    \n",
    "    h, w = img.shape[:2]\n",
    "\n",
    "    # width and height of the zoomed image\n",
    "    zh = int(np.round(zoom_factor * h))\n",
    "    zw = int(np.round(zoom_factor * w))\n",
    "\n",
    "    # for multichannel images we don't want to apply the zoom factor to the RGB\n",
    "    # dimension, so instead we create a tuple of zoom factors, one per array\n",
    "    # dimension, with 1's for any trailing dimensions after the width and height.\n",
    "    zoom_tuple = (zoom_factor,) * 2 + (1,) * (img.ndim - 2)\n",
    "\n",
    "    # zooming out\n",
    "    if zoom_factor < 1:\n",
    "        # bounding box of the clip region within the output array\n",
    "        top = (h - zh) // 2\n",
    "        left = (w - zw) // 2\n",
    "        # zero-padding\n",
    "        out = np.zeros_like(img)\n",
    "        out[top:top+zh, left:left+zw] = ndimage.interpolation.zoom(img, zoom_tuple, **kwargs)\n",
    "\n",
    "    # zooming in\n",
    "    elif zoom_factor > 1:\n",
    "        # bounding box of the clip region within the input array\n",
    "        top = (zh - h) // 2\n",
    "        left = (zw - w) // 2\n",
    "        out = ndimage.interpolation.zoom(img[top:top+zh, left:left+zw], zoom_tuple, **kwargs)\n",
    "        # `out` might still be slightly larger than `img` due to rounding, so\n",
    "        # trim off any extra pixels at the edges\n",
    "        trim_top = ((out.shape[0] - h) // 2)\n",
    "        trim_left = ((out.shape[1] - w) // 2)\n",
    "        out = out[trim_top:trim_top+h, trim_left:trim_left+w]\n",
    "\n",
    "    # if zoom_factor == 1, just return the input array\n",
    "    else:\n",
    "        out = img\n",
    "    return out\n",
    "\n",
    "def generate_more_images(img_set, actual_labels):\n",
    "    \n",
    "    ## checking that the inputs are the correct lengths\n",
    "    assert img_set.shape[0] == len(actual_labels)\n",
    "    \n",
    "    new_img_set = []\n",
    "    new_y = []\n",
    "    \n",
    "    ## count the number of occurrences in each class and find the avg class size\n",
    "    class_counts = np.bincount(actual_labels)\n",
    "    avg_class_size = np.round(np.mean(class_counts))\n",
    "    \n",
    "    ## iterate through each class to see how many images need to be added\n",
    "    for class_num in tqdm(range(len(class_counts))):\n",
    "        \n",
    "        img_idx = np.where(actual_labels == class_num)\n",
    "        \n",
    "        if class_counts[class_num] < avg_class_size:\n",
    "            num_to_add = np.int(avg_class_size - class_counts[class_num])\n",
    "            \n",
    "            ## randomly assign a transformation (0.85 because zoom operation takes forever)\n",
    "            for _ in tqdm(range(num_to_add)):\n",
    "                if random.random() < 0.85:\n",
    "                    new_img_set.append(rotate_image(img_set[random.choice(img_idx)])[0:32,0:32,:])\n",
    "                    new_y.append(class_num)\n",
    "                else:\n",
    "                    new_img_set.append(clipped_zoom(img_set[random.choice(img_idx)])[0:32,0:32,:])\n",
    "                    new_y.append(class_num)\n",
    "    \n",
    "    return np.array(new_img_set), np.array(new_y)\n",
    "                    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 686,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "## generate more images\n",
    "#new_x, new_y = generate_more_images(X_train, y_train)\n",
    "\n",
    "## save the generated images\n",
    "additional_data_file = '/Users/jwen/Python/Projects/sdcnd_traffic_sign/traffic-signs-data/additional_data.p'\n",
    "\n",
    "with open(additional_data_file, 'wb') as handle:\n",
    "    pickle.dump({\"features\":new_x, \"labels\":new_y}, handle, protocol=pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 684,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "## fix the shape of the new image array\n",
    "# new_x = np.array([i[0] for i in new_x])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 715,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x13876add8>"
      ]
     },
     "execution_count": 715,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATAAAAE8CAYAAABHOe2IAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAMKigAEIiwIGICwKGICw/i86Dm4tYMs2wwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13883c630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## print newly generated image to test and show\n",
    "index = 5432\n",
    "plt.figure(figsize=(5,5))\n",
    "plt.grid(False)\n",
    "plt.title(str(sign_names[sign_names[\"ClassId\"] == new_y[index]].iloc[0,1]) + \" - Class \" + str(new_y[index]))\n",
    "plt.tick_params(axis='x', colors=(0,0,0,0))\n",
    "plt.tick_params(axis='y', colors=(0,0,0,0))\n",
    "plt.imshow(new_x[index])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 724,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "## combine the generated images and the original training set\n",
    "X_train = np.append(X_train, new_x, axis=0)\n",
    "y_train = np.append(y_train, new_y, axis=0)\n",
    "\n",
    "## preprocess data\n",
    "X_train = preprocess(X_train)[:,:,:,np.newaxis]\n",
    "# X_valid = preprocess(X_valid)[:,:,:,np.newaxis]\n",
    "\n",
    "## save the preprocessed training images\n",
    "train_data_all_file = '/Users/jwen/Python/Projects/sdcnd_traffic_sign/traffic-signs-data/train_data_all.p'\n",
    "\n",
    "with open(train_data_all_file, 'wb') as handle:\n",
    "    pickle.dump({\"features\":X_train, \"labels\":y_train}, handle, protocol=pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 730,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(11915, 32, 32, 3)\n",
      "(46714, 32, 32, 1)\n",
      "Added 11915 images\n"
     ]
    }
   ],
   "source": [
    "## added 11915 images\n",
    "print(new_x.shape)\n",
    "print(X_train.shape)\n",
    "print(\"Added {} images\".format(new_x.shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Reloading the pickled training data onto server\n",
    "import pickle\n",
    "\n",
    "## bringing in the train, validation, and test data\n",
    "training_file = '/home/carnd/sdcnd_traffic_sign/traffic-signs-data/train_data_all.p'\n",
    "validation_file= '/home/carnd/sdcnd_traffic_sign/traffic-signs-data/valid.p'\n",
    "testing_file = '/home/carnd/sdcnd_traffic_sign/traffic-signs-data/test.p'\n",
    "\n",
    "with open(training_file, mode='rb') as f:\n",
    "    train = pickle.load(f)\n",
    "with open(validation_file, mode='rb') as f:\n",
    "    valid = pickle.load(f)\n",
    "with open(testing_file, mode='rb') as f:\n",
    "    test = pickle.load(f)\n",
    "    \n",
    "X_train, y_train = train['features'], train['labels']\n",
    "X_valid, y_valid = valid['features'], valid['labels']\n",
    "X_test, y_test = test['features'], test['labels']\n",
    "\n",
    "## need to preprocess this\n",
    "X_valid = preprocess(X_valid)[:,:,:,np.newaxis]\n",
    "X_test = preprocess(X_test)[:,:,:,np.newaxis]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(46714, 32, 32, 1)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Model Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.contrib.layers import flatten\n",
    "\n",
    "### Define your architecture\n",
    "def conv2d(x, weights, biases, strides=1, padding='VALID'):\n",
    "    conv = tf.nn.conv2d(x, weights, strides=[1, strides, strides, 1], padding=padding)\n",
    "    conv = tf.add(conv, biases)\n",
    "    return conv\n",
    "\n",
    "def maxpool2d(x, k=2, padding='VALID'):\n",
    "    return tf.nn.max_pool(x, ksize=[1, k, k, 1], strides=[1, k, k, 1], padding=padding)\n",
    "\n",
    "def LeNet(x):    \n",
    "    # Arguments used for tf.truncated_normal, randomly defines variables for the weights and biases for each layer\n",
    "    mu = 0\n",
    "    sigma = 0.1\n",
    "    \n",
    "    weights = {\n",
    "        'c1': tf.Variable(tf.truncated_normal([5, 5, 1, 6], mean=mu, stddev=sigma), name='c1_w'),\n",
    "        'c2': tf.Variable(tf.truncated_normal([5, 5, 6, 16], mean=mu, stddev=sigma), name='c2_w'),\n",
    "        'fc1': tf.Variable(tf.truncated_normal([400, 120], mean=mu, stddev=sigma), name='fc1_w'),\n",
    "        'fc2': tf.Variable(tf.truncated_normal([120, 84], mean=mu, stddev=sigma), name='fc2_w'),\n",
    "        'fc3': tf.Variable(tf.truncated_normal([84, 43], mean=mu, stddev=sigma), name='fc3_w'),\n",
    "    }\n",
    "    \n",
    "    biases = {\n",
    "        'c1': tf.Variable(tf.zeros(6), name='c1_b'),\n",
    "        'c2': tf.Variable(tf.zeros(16), name='c2_b'),\n",
    "        'fc1': tf.Variable(tf.zeros(120), name='fc1_b'),\n",
    "        'fc2': tf.Variable(tf.zeros(84), name='fc2_b'),\n",
    "        'fc3': tf.Variable(tf.zeros(43), name='fc3_b'),\n",
    "    }\n",
    "    \n",
    "    # Layer 1\n",
    "    conv1 = conv2d(x, weights['c1'], biases['c1'], strides = 1) # Convolutional. Input = 32x32x3. Output = 28x28x6.\n",
    "    conv1 = tf.nn.relu(conv1) # Activation.\n",
    "    conv1 = maxpool2d(conv1) # Pooling. Input = 28x28x6. Output = 14x14x6.\n",
    "\n",
    "    # Layer 2\n",
    "    conv2 = conv2d(conv1, weights['c2'], biases['c2'], strides = 1) # Convolutional. Output = 10x10x16.\n",
    "    conv2 = tf.nn.relu(conv2) # Activation.\n",
    "    conv2 = maxpool2d(conv2) # Pooling. Input = 10x10x16. Output = 5x5x16.\n",
    "\n",
    "    # Flatten\n",
    "    flatten_layer = flatten(conv2) # Flatten. Input = 5x5x16. Output = 400.\n",
    "    \n",
    "    # Layer 3\n",
    "    fc1 = tf.add(tf.matmul(flatten_layer, weights['fc1']), biases['fc1']) # Fully Connected. Input = 400. Output = 120.\n",
    "    fc1 = tf.nn.relu(fc1) # Activation.\n",
    "    fc1 = tf.nn.dropout(fc1, keep_prob) ## Dropout\n",
    "\n",
    "    # Layer 4\n",
    "    fc2 = tf.add(tf.matmul(fc1, weights['fc2']), biases['fc2']) # Fully Connected. Input = 120. Output = 84.\n",
    "    fc2 = tf.nn.relu(fc2) # Activation.\n",
    "    fc2 = tf.nn.dropout(fc2, keep_prob) ## Dropout\n",
    "\n",
    "    # Layer 5\n",
    "    fc3 = tf.add(tf.matmul(fc2, weights['fc3']), biases['fc3']) # Fully Connected. Input = 84. Output = 43.\n",
    "    \n",
    "    logits = fc3\n",
    "    \n",
    "    return logits"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Train, Validate and Test the Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "A validation set can be used to assess how well the model is performing. A low accuracy on the training and validation\n",
    "sets imply underfitting. A high accuracy on the training set but low accuracy on the validation set implies overfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "## Placeholders for the x and y variables\n",
    "x = tf.placeholder(tf.float32, (None, 32, 32, 1))\n",
    "y = tf.placeholder(tf.int32, (None))\n",
    "keep_prob = tf.placeholder(tf.float32)\n",
    "one_hot_y = tf.one_hot(y, 43)\n",
    "\n",
    "## Parameters\n",
    "RATE = 0.001\n",
    "EPOCHS = 25\n",
    "BATCH_SIZE = 128\n",
    "\n",
    "## Model pipeline\n",
    "logits = LeNet(x)\n",
    "cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits, one_hot_y)\n",
    "loss_operation = tf.reduce_mean(cross_entropy)\n",
    "optimizer = tf.train.AdamOptimizer(learning_rate = RATE)\n",
    "training_operation = optimizer.minimize(loss_operation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Setting up model evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "## Calculate model accuracy\n",
    "correct_prediction = tf.equal(tf.argmax(logits, 1), tf.argmax(one_hot_y, 1))\n",
    "accuracy_operation = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
    "saver = tf.train.Saver()\n",
    "\n",
    "def evaluate(X_data, y_data):\n",
    "    num_examples = len(X_data)\n",
    "    total_accuracy = 0\n",
    "    sess = tf.get_default_session()\n",
    "    for offset in range(0, num_examples, BATCH_SIZE):\n",
    "        batch_x, batch_y = X_data[offset:offset+BATCH_SIZE], y_data[offset:offset+BATCH_SIZE]\n",
    "        accuracy = sess.run(accuracy_operation, feed_dict={x: batch_x, y: batch_y, keep_prob: 1.0})\n",
    "        total_accuracy += (accuracy * len(batch_x))\n",
    "    return total_accuracy / num_examples"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Running model training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training...\n",
      "\n",
      "EPOCH 1 ...\n",
      "Training Accuracy = 0.858\n",
      "Validation Accuracy = 0.801\n",
      "\n",
      "EPOCH 2 ...\n",
      "Training Accuracy = 0.939\n",
      "Validation Accuracy = 0.868\n",
      "\n",
      "EPOCH 3 ...\n",
      "Training Accuracy = 0.965\n",
      "Validation Accuracy = 0.886\n",
      "\n",
      "EPOCH 4 ...\n",
      "Training Accuracy = 0.975\n",
      "Validation Accuracy = 0.902\n",
      "\n",
      "EPOCH 5 ...\n",
      "Training Accuracy = 0.980\n",
      "Validation Accuracy = 0.917\n",
      "\n",
      "EPOCH 6 ...\n",
      "Training Accuracy = 0.984\n",
      "Validation Accuracy = 0.925\n",
      "\n",
      "EPOCH 7 ...\n",
      "Training Accuracy = 0.985\n",
      "Validation Accuracy = 0.925\n",
      "\n",
      "EPOCH 8 ...\n",
      "Training Accuracy = 0.990\n",
      "Validation Accuracy = 0.934\n",
      "\n",
      "EPOCH 9 ...\n",
      "Training Accuracy = 0.990\n",
      "Validation Accuracy = 0.934\n",
      "\n",
      "EPOCH 10 ...\n",
      "Training Accuracy = 0.992\n",
      "Validation Accuracy = 0.941\n",
      "\n",
      "EPOCH 11 ...\n",
      "Training Accuracy = 0.994\n",
      "Validation Accuracy = 0.941\n",
      "\n",
      "EPOCH 12 ...\n",
      "Training Accuracy = 0.994\n",
      "Validation Accuracy = 0.948\n",
      "\n",
      "EPOCH 13 ...\n",
      "Training Accuracy = 0.993\n",
      "Validation Accuracy = 0.940\n",
      "\n",
      "EPOCH 14 ...\n",
      "Training Accuracy = 0.996\n",
      "Validation Accuracy = 0.942\n",
      "\n",
      "EPOCH 15 ...\n",
      "Training Accuracy = 0.996\n",
      "Validation Accuracy = 0.948\n",
      "\n",
      "EPOCH 16 ...\n",
      "Training Accuracy = 0.996\n",
      "Validation Accuracy = 0.947\n",
      "\n",
      "EPOCH 17 ...\n",
      "Training Accuracy = 0.997\n",
      "Validation Accuracy = 0.949\n",
      "\n",
      "EPOCH 18 ...\n",
      "Training Accuracy = 0.997\n",
      "Validation Accuracy = 0.950\n",
      "\n",
      "EPOCH 19 ...\n",
      "Training Accuracy = 0.997\n",
      "Validation Accuracy = 0.949\n",
      "\n",
      "EPOCH 20 ...\n",
      "Training Accuracy = 0.997\n",
      "Validation Accuracy = 0.951\n",
      "\n",
      "EPOCH 21 ...\n",
      "Training Accuracy = 0.998\n",
      "Validation Accuracy = 0.952\n",
      "\n",
      "EPOCH 22 ...\n",
      "Training Accuracy = 0.997\n",
      "Validation Accuracy = 0.946\n",
      "\n",
      "EPOCH 23 ...\n",
      "Training Accuracy = 0.998\n",
      "Validation Accuracy = 0.953\n",
      "\n",
      "EPOCH 24 ...\n",
      "Training Accuracy = 0.997\n",
      "Validation Accuracy = 0.951\n",
      "\n",
      "EPOCH 25 ...\n",
      "Training Accuracy = 0.998\n",
      "Validation Accuracy = 0.950\n",
      "\n",
      "Model saved\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    num_examples = len(X_train)\n",
    "    \n",
    "    print(\"Training...\")\n",
    "    print()\n",
    "    \n",
    "    ## store tore the performance for plots\n",
    "    training_acc_hist = []\n",
    "    validation_acc_hist = []\n",
    "    \n",
    "    for i in range(EPOCHS):\n",
    "        X_train, y_train = shuffle(X_train, y_train)\n",
    "        for offset in range(0, num_examples, BATCH_SIZE):\n",
    "            end = offset + BATCH_SIZE\n",
    "            batch_x, batch_y = X_train[offset:end], y_train[offset:end]\n",
    "            sess.run(training_operation, feed_dict={x: batch_x, y: batch_y, keep_prob:0.6})\n",
    "        \n",
    "        ## compute the accuracy for the training/ validation sets\n",
    "        training_accuracy = sess.run(accuracy_operation, feed_dict={x: X_train, y: y_train, keep_prob: 1.0})\n",
    "        validation_accuracy = sess.run(accuracy_operation, feed_dict={x: X_valid, y: y_valid, keep_prob: 1.0})\n",
    "        \n",
    "        ## store history for plotting curves\n",
    "        training_acc_hist.append(training_accuracy)\n",
    "        validation_acc_hist.append(validation_accuracy)\n",
    "        \n",
    "        validation_predictions = tf.argmax(logits, 1)\n",
    "        validation_logits = logits\n",
    "        \n",
    "        print(\"EPOCH {} ...\".format(i+1))\n",
    "        print(\"Training Accuracy = {:.3f}\".format(training_accuracy))\n",
    "        print(\"Validation Accuracy = {:.3f}\".format(validation_accuracy))\n",
    "        print()\n",
    "        \n",
    "    saver.save(sess, './lenet_traffic_aws')\n",
    "    print(\"Model saved\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validation Set Performance Metrics\n",
    "\n",
    "Taking a look at the performance of the model on the validation set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Validation Accuracy = 0.950\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "\n",
    "    validation_accuracy = sess.run(accuracy_operation, feed_dict={x: X_valid, y: y_valid, keep_prob: 1.0})\n",
    "    print(\"Validation Accuracy = {:.3f}\".format(validation_accuracy))\n",
    "    validation_pred_label = sess.run([tf.argmax(logits, 1)], feed_dict={x: X_valid, y: y_valid, keep_prob: 1.0})\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7f68a14a7e48>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/carnd/anaconda3/envs/carnd-term1/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3sAAAG5CAYAAAA3ci11AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJBXIYk+SJEmSCvT/AeucjfHnLUspAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f68c9d26390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,7))\n",
    "bins = np.arange(44) - 0.5\n",
    "plt.hist([y_valid, validation_pred_label[0]], label = ['Actual','Predicted'], bins=bins)\n",
    "plt.legend()\n",
    "plt.xticks(range(43))\n",
    "plt.title(\"Actual vs. Predicted Histogram\")\n",
    "plt.xlabel('Sign Index')\n",
    "plt.ylabel('Count')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f68a1ff9a90>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/carnd/anaconda3/envs/carnd-term1/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAk8AAAJcCAYAAADghUBbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xe4ZVWVqP23AlGRnJrCBhUGKFIlIpcLjUhQURFERUGk\nySawBRMiPqC329vaKEjbtl5JYosIIrRegspnq2g3goIkLQYm0IJqECUpEip8f6x1vIfjTuvUCucc\n3t/z1FM7zLnn2HvuMM5cYcxavnw5kiRJGs3srgOQJEmaTkyeJEmSKjB5kiRJqsDkSZIkqQKTJ0mS\npApMniRJkioweZLUuohYLSL+b0Q8EBFfXoHHOSgivllnbF2IiCsi4pCu45A0mlme50lSPxHxBuCd\nwFbAQ8ANwIcz8/sr+LgHA28HdsrMJSscaM0i4kXAt4FLMvPV426fT/EafDczXzTC43wQeFZmvrGZ\nSCV1wZUnST1FxDuBTwD/G9gQeDrwr8C+NTz8XwO3TcXEaZzfAjtFxLrjbjsEuK2uASJiVkT4PSxN\nM648SfoLEbEmcCdwWGb23KwWEasAHwVeV950IXB8Zj5artx8ATgNOB5YCrw/M8+JiA8BJwCzgEeB\ndwCbMm6FJiI2A34FrJSZSyLiUOAkYH3gXuADmXleefuRmfk3Zb+dgNOBLSmSnHdk5n+V930H+B6w\nO7AtcDXwhsy8t8dzG4v/UuDmzPxURMwB7gA+C+w+tvIUEacDrwbWBH4GHJuZ34uIvYCvjXuev8jM\n+WUc/wm8CNgOeC5wJvCFzDwzIj4NrJ+Zry0f/6PA9sCemekXtjQF+BePpF7+J7AqcMmANicCOwIL\ngPnADsAHxt2/EUVCsQlwBPCpiFg7M0+mWM26IDOfmplnDQokIp4C/DPwssxcA9iJYtPZxHbrAJeV\nbdcFTgUum7By9AbgMGADYGXg3YPGBj4P/G15+aXAT4C7JrT5IcVrsA7wReDLEbFqZn59wvOcP67P\nwcCbgDUoErLx3gVsGxGHRsQuFK/dISZO0tRh8iSpl3WBe4dsVjsI+F+ZeU9m/hb4EEVSMObx8v7H\nM/Ny4A9ATDKeZcA2EbFaZi7OzJ/0aPMK4GeZ+W+ZuSQzzwduBV45rs05mXlbZv6JYqVswaBBy1Wr\ndSIiKJKoz/do84XM/F055seBVRj+PD+XmT8p+zw+4fEeBt5Ikfx9AXh7Zi4a8niSWmTyJKmX3wHr\nRcTcAW3+iieumtxR3vbnx5iQfD0MPLVqIJn5R+D1wFuAxRFxWURsNUI8YzFtMu76f08inn8DjgF2\no8dKXES8KyIWlkcO3k+x2rbekMf8zaA7M/Na4JcUm/wuHCFGSS0yeZLUy9XAI8CrBrS5i2LH7zFP\n5y83aY3qj8Dq465vNP7OzPxGZr4Y2JhiNemMEeIZi+nOScY05t+AtwGXl6tCf1ZuVjueYr+vtTNz\nLeABiqQHoN+mtoGb4CLiaIoVrLuA904+dElNGPRXpaQnqcx8ICJOothPaQnwTYrNcHsCu2Xme4Hz\ngQ9ExA8pkoGTKDYzTcYNwPER8XSK5OOEsTsiYkPgfwDfAv5EsflvaY/HuBz4ZHl6hQuB1wDPptjp\ne9Iy81cRsSvFStBEawBLKI7MmxsR7wOeNu7+u4EXR8TszFw2yngRsSXwDxQ7lD8MXBsRV2TmX+zn\nJakbrjxJ6ikzT6U4x9MHKJKD31Bsvvr3ssk/AD8CbgJuBq4vb5vMWFcCF5SPdR1PTHhmU+xEfRfw\ne2BXipWgiY/xO2Dvsu3vKFZs9u51NN0k4vt+ZvZaVfsGcAXFkX13UKzWjd8kN3ak4u8i4vph45Sb\nSb8AfDQzb8zMnwHvB/6tPLpR0hTgqQokSZIqcOVJkiSpApMnSZKkCkyeJEmSKjB5kiRJqmBanKrg\n7gcfr7xX+5qrr9REKCvsT4/1OsK6v9VWntNQJCtmMgcazJo1a3gjSZKmiFXn0vOHy5UnSZKkCkye\nJEmSKjB5kiRJqqCTfZ4iYi/gdGAOcGZmfqSLOCRJkqpqfeUpIuYAnwJeRlF36sCIeHbbcUiSJE1G\nF5vtdgB+npm/zMzHgC8B+3YQhyRJUmVdJE+b8MTCmYvK2yRJkqa8LpKnXudMsDqxJEmaFrpInhYB\nm467Pg+4q4M4JEmSKuviaLsfAltExObAncABwBs6iEOSJKmy1leeMnMJcAzwDWAhcGFm/qTtOCRJ\nkiZj1mRqlLXN2nZTj7XtJEkzXb/adtMieXpkSfUdytfe8+8rtb/76ydWHYKV51ZfuFuydFml9nNm\nV084TFJmhqqfTeddUhPa+C6aqn+QWxhYkiSpBiZPkiRJFXRV2+5sYG/gnszcposYJEmSJqOrlafP\nAXt1NLYkSdKkdZI8ZeZVwO+7GFuSJGlFuM+TJElSBSZPkiRJFZg8SZIkVWDyJEmSVEEnyVNEnA9c\nXVyMRRFxRBdxSJIkVdXJeZ4y88AuxpUkSVpRM7a23dJl1bqst9sHqg7Bfd/9cOU+VV/vqs8DYO6c\n5hcUq9bog3bikiSpLta2kyRJqoHJkyRJUgUmT5IkSRW0vsN4RGwKfB7YCFgGfDYzT287DkmSpMno\nYuVpCfCuzNwa2BE4OiKe3UEckiRJlbWePGXm4sy8vrz8ELAQ2KTtOCRJkiaj032eImIz4HnANV3G\nIUmSNKrOkqeIeCrwFeDYzHywqzgkSZKq6Ko8y0oUidN5mXlxFzFIkiRNRuvJU0TMAs4CFmbmqW2P\nL0mStCK6qG23M3AwcHNE3FDe9v7MvLyDWCRJkippPXnKzO9D71oxkiRJU92MLQzchld8+urKfS59\ny46V2s+aZZ45E0zmc+bcS1K3LAwsSZJUA5MnSZKkCrqobbcqcBWwSjn+RZl5cttxSJIkTUYXK0+P\nArtn5nxgAbBXRFTbEUiSJKkjXRxttxz4Q3l1pfLflNwhXJIkaaIuzvNERMwBrgOeBXwqM61tJ0mS\npoVOdhjPzKWZuQCYB+wQEdt0EYckSVJVnR5tl5n3A98B9uoyDkmSpFF1Udtu/YhYq7y8GrAncGvb\ncUiSJE1GF/s8bQycW+73NBu4MDMv7SAOSZKkyro42u4m4HltjytJklQHzzAuSZJUgYWBW7b2az9b\nqf19F72poUgkSdIgFgaWJEmqgcmTJElSBZ2cYRz+fJbxHwF3ZubeXcUhSZJURZcrT+8AFnY4viRJ\nUmWdJE8RMQ94BXBmF+NLkiRNVlcrT58A3gss62h8SZKkSemiPMvewD2ZeV3bY0uSJK2oLlaedgb2\niYjbgS8Bu0fEFzqIQ5IkqbIuyrOcAJwAEBEvAt6dmW9sOw5JkqTJ8DxPkiRJFXR2nieAzPwO8J0u\nY5AkSapixta2q/q8liyt/jqsNLf5hbsN//bfKve5+/MHNxDJzFX1vTJrVs9SR1ItJvOd7HtS091U\nfd9b206SJKkGJk+SJEkVmDxJkiRV0MkO4+U5nh4ClgJLMnP7LuKQJEmqqsuj7XbLzHs7HF+SJKky\nN9tJkiRV0FXytBz4ZkRcFxFv6igGSZKkyrpKnnbOzO2AlwFHR8QLO4pDkiSpkk6Sp8y8q/z/HuAS\nYIcu4pAkSaqq9eQpIp4SEWuMXQZeAtzSdhySJEmT0cXRdhsCl0TE2PhfzMyvdxCHJElSZa0nT5n5\nS2B+2+NKkiTVYcYWBn4yu+H2+yv3WbDZWg1EIklTz7Jl1X5SZs+28PKTlYWBJUmSamDyJEmSVEFX\nte3WAs4EtqE4YebhmXl1F7FIkiRV0dXK0+nA1zNzK4qdxxd2FIckSVIlra88RcTTgBcChwJk5mPA\nY23HIUmSNBldbLZ7BvBb4JyImA9cB7wjM//YQSySJEmVdLHZbi6wHfDpzHwe8EfgfR3EIUmSVFkX\nydMiYFFmXlNev4gimZIkSZryWk+eMvO/gd9EWZ8F2AP4adtxSJIkTUYnpyoA3g6cFxErA78EDuso\nDkmSpEo6SZ4y8wZg+y7GliRJWhGeYVySJKkCCwMLgP3OuGZ4o3EuPOwFlcdYaa65uiRp+rAwsCRJ\nUg1MniRJkiroojxLABeMu+kZwEmZ+Ym2Y5EkSaqq9eQpMxNYABARc4A7gUvajkOSJGkyut5stwfw\ni8y8o+M4JEmSRtJ18nQAcH7HMUiSJI2ss+SpPLv4PsCXu4pBkiSpqi5Xnl4GXJ+Zd3cYgyRJUiVd\nJk8H4iY7SZI0zXSSPEXE6sCLgYu7GF+SJGmyuioM/DCwbhdjS5IkrYhpUdvu4ceqBzl7ds9yNLWa\nzGs3a1bzcU3GsmXVnstrzrq28hiXHPU/KveRpLpV/e6eqt/bap617SRJkmpg8iRJklSByZMkSVIF\nnewwHhHHAUcCy4GbgcMy85EuYpEkSaqi9ZWniNgE+Dtg+8zcBphDUaZFkiRpyutqs91cYLWImAus\nDtzVURySJEmVtJ48ZeadwMeAXwOLgQcy85ttxyFJkjQZXWy2WxvYF9gc+CvgKRHxxrbjkCRJmowu\nNtvtCfwqM3+bmY9TlGjZqYM4JEmSKuviaLtfAzuW9e3+BOwB/KiDOCRJkirrYp+na4CLgOspTlMw\nG/hs23FIkiRNRleFgU8GTu5ibEmSpBUxLQoDP7KEqR/kk8xk3jev+PTVlftc/jZ3h5MkdcPCwJIk\nSTUweZIkSaqgq9p27wCOAmYBZ2TmJ7qIQ5IkqaouTpK5DUXitAMwH9g7IrZoOw5JkqTJ6GKz3dbA\nDzLz4cxcAnwX2K+DOCRJkirrYrPdLcCHI2JdipNkvhxPkilJkqaJLk6SuRD4KHAl8HXgRmBJ23FI\nkiRNRlcnyTwLOAsgIv43sKiLOCRJkqrq5FQFEbFB+f/TgVcD53cRhyRJUlWdrDwBXyn3eXocODoz\n7+soDkmSpEq62my3SxfjSpIkrShr22lKW/u1n63U/r6L3tRQJJKkJxtr20mSJNXA5EmSJKkCkydJ\nkqQKGtthPCLOBvYG7snMbcrb1gEuADYDbgde55F2kiRpOmly5elzwF4Tbnsf8K3M3AL4VnldkiRp\n2mgsecrMq4DfT7h5X+Dc8vK5wKuaGl+SJKkJbe/ztGFmLgYo/9+g5fElSZJWiDuMS5IkVdB28nR3\nRGwMUP5/T8vjS5IkrZC2k6evAYeUlw8Bvtry+JIkSSukyVMVnA+8CFgvIhYBJwMfAS6MiCOAXwP7\nNzW+JElSExpLnjLzwD537dHUmJIkSU2zMLBmlJf/639V7nP523ZqIBJJk1H1N2nWrJ51W6VaWBhY\nkiSpBiZPkiRJFZg8SZIkVdB2YeD9gQ8CWwM7ZOaPmhpfkiSpCW0XBr4FeDVwVYPjSpIkNabVwsCZ\nuTAzs6kxJUmSmuY+T5IkSRWYPEmSJFVg8iRJklSByZMkSVIFbRcG/j3wSWB94LKIuCEzX9pUDJIk\nSXXrojDwJU2NKUmS1DQLA5cm8zpYkHJmWHv/Myq1v+/LRzUUiSRpKrEwsCRJUg1MniRJkipou7bd\nKcArgceAXwCHZeb9TcUgSZJUt7Zr210JbJOZ2wK3ASc0OL4kSVLt2q5t983MXFJe/QEwr6nxJUmS\nmtDlPk+HA1d0OL4kSVJlnSRPEXEisAQ4r4vxJUmSJquxHcb7iYhDKHYk3yMzp/5JpiRJksZpNXmK\niL2A44FdM/PhNseWJEmqQ9u17U4AVgGujAiAH2TmW5qKQZIkqW5t17Y7q6nxJEmS2mBtO6mitV9w\nTOU+9/3wXxqIRNJMUvX3eCbVV52q9WWtbSdJklQDkydJkqQKTJ4kSZIqaLsw8N8D+wLLgHuAQzPz\nrqZikCRJqlvbhYFPycxtM3MBcClwUoPjS5Ik1a7twsAPjrv6FPAoOkmSNL10UZ7lw8DfAg8Au7U9\nviRJ0opofYfxzDwxMzelKApc/YQ5kiRJHeryaLsvAq/pcHxJkqTKWk2eImKLcVf3AW5tc3xJkqQV\n1XZh4JdHURF4GXAHYFFgSZI0rVgYWJIkqQILA0st+OL1d1Tu84bt/rqBSCRJo7IwsCRJUg1MniRJ\nkioweZIkSaqg1cLA4+57N3AKsH5m3ttUDJIkSXVruzAwEbEp8GLg1w2OLUmS1IhWCwOXTgPei0WB\nJUnSNNT2Gcb3Ae7MzBvbHFeSJKkuje3zNFFErA6cCLykrTElSZLq1ubK0zOBzYEbI+J2YB5wfURs\n1GIMkiRJK6S1lafMvBnYYOx6mUBt79F2kiRpOmls5aksDHx1cTEWRcQRTY0lSZLUlrYLA4+/f7Om\nxpYkSWqKhYGlKapqMWELCUtSvSwMLEmSVAOTJ0mSpAparW0XER8EjgJ+WzZ7f2Ze3lQMkiRJdWvy\nVAWfA/4F+PyE20/LzI81OK4kSVJjuqhtJ0mSNG11sc/TMRFxU0ScHRFrdzC+JEnSpLWdPH2aokzL\nAmAx8PGWx5ckSVohrZVnAcjMu8cuR8QZwKVtji9JkrSiWl15ioiNx13dD7ilzfElSZJWVJOnKjgf\neBGwXkQsAk4GXhQRC4DlwO3Am5saX5IkqQlt17Y7q6nxJEmS2mBtO2mG+PHt91fu87zN1mogEklq\nXtX8ZdasnmXqBrK2nSRJUg1MniRJkioweZIkSaqg1cLA5e1vB44BlgCXZeZ7m4pBkiSpbk2uPH0O\n2Gv8DRGxG7AvsG1mPgewQLAkSZpW2i4M/FbgI5n5aNnmnqbGlyRJakLb+zxtCewSEddExHcj4gUt\njy9JkrRC2k6e5gJrAzsC7wEujIjqJ16QJEnqSNvJ0yLg4sxcnpnXAsuA9VqOQZIkadLaTp7+Hdgd\nICK2BFYG7m05BkmSpElruzDw2cDZEXEL8BhwSGZaekWSJE0b1raTZghr20l6Mumytp3Jk9SCJUuX\nVe4zd07zW9VPuPzWyn3+8eVbNRCJJE09FgaWJEmqgcmTJElSBa3WtouIC4Aom6wF3J+ZC5qKQZIk\nqW6NJU8Ute3+Bfj82A2Z+fqxyxHxceCBBseXJEmqXdu17QAozyr+OuD8psaXJElqQlf7PO0C3J2Z\nP+tofEmSpEnpKnk6EFedJEnSNNTkPk89RcRc4NXA89seW5IkaUV1sfK0J3BrZi7qYGxJkqQV0ljy\nVNa2u7q4GIsi4ojyrgNwk50kSZqmGttsl5kH9rn90KbGlCRJappnGJckSarAwsCa0tqomq1qPnBF\ntWLC//AyCwlXMZnv5GWT+IacXfGj4mdLT0YWBpYkSaqByZMkSVIFbRcGXgB8BlgVWAK8LTOvbSoG\nSZKkujW58vQ5YK8Jt/0T8KHMXACcVF6XJEmaNtouDLwceFp5eU3grqbGlyRJakLb5VmOBb4RER+j\nSNx2anl8SZKkFdL2DuNvBY7LzE2B44CzWh5fkiRphbSdPB0CXFxe/jKwQ8vjS5IkrZC2k6e7gF3L\ny7sDP2t5fEmSpBXS5KkKzgdeBKwXEYuAk4GjgNMjYi7wCPCmpsaXJElqQuuFgYHnNzWmJElS06xt\npyltWcWiXbOrFuxS4+5+4JHKfTZcc9UGItFMNJnfMOv0aVTWtpMkSaqByZMkSVIFJk+SJEkVtF0Y\neD5FYeCnArcDB2Xmg03FIEmSVLe2CwOfCbwvM58LXAK8p8HxJUmSatd2YeAAriovXwm8pqnxJUmS\nmtD2Pk+3APuUl/cHNm15fEmSpBXSdvJ0OHB0RFwHrAE81vL4kiRJK6SxHcZ7ycxbgZcARMSWwCva\nHF+SJGlFtbryFBEblP/PBj5AceSdJEnStNF2YeCnRsTRZZOLgXOaGl+SJKkJ1rbTlGZtu+nP2nZq\nkrXt1KR+te1MnvSkV/Uz4Bdv825b/FDlPltuvEal9pP57quYywMwp2JC39Z3ctX3cVtJypKlyyq1\nnzvHQhkzwVRNgi0MLEmSVAOTJ0mSpAqa3GF8U+DzwEbAMuCzmXl6RKwDXABsRlHf7nWZeV9TcUiS\nJNWpyZWnJcC7MnNrYEeKk2M+G3gf8K3M3AL4VnldkiRpWui78hQRX4b+O2pn5usGPXBmLgYWl5cf\nioiFwCbAvhSnMAA4F/gOcHyVoCVJkroyaLPdpXUNEhGbAc8DrgE2LBMrMnPx2IkzJUmSpoO+yVNm\nnlvHABHxVOArwLGZ+WBE1PGwkiRJnRi6w3hEzKUo6LsA+POZ6zLz8BH6rkSROJ2XmReXN98dERuX\nq04bA/dMKnJJkqQOjLLD+P8Bdgb2Bn4GvAD407BOETELOAtYmJmnjrvra8Ah5eVDgK9WCViSJKlL\noyRPO2TmIcD9mfmPwN8Azxmh387AwcDuEXFD+e/lwEeAF0fEz4AXl9clSZKmhVHO8zS2yrQ0IlbP\nzAdG2ck7M78PvU9rDuwxaoCSJElTySjJ0+8jYm3g68AVEXEvcGezYUmSJE1NoyRPr8jMpRFxInAQ\nsCbFmcOlKadqUVGwsGhVSytWx61aGBeqF/kF+M3vHq7UftN1V688xqwWivZWfX0BZk+qQOrULIhd\n9f3y2JLqn/mV5/qZr6KN4unTreD6rLYqeK+IR5ZU/JTrScvkqXltJE+T0UbytGwyiU3F5z+Z9/Bk\nkqeqXdr6cav6m/T40upzYvJUTRvJ01S16tzeux+NcqqC39LjT5TM9OSWkiTpSWeUzXbbj7u8KsWm\nu8eHdRpQGHh/4IPA1hRH8v2oatCSJEldGZo8ZeYdE246KSJ+APz9kK5jhYGvj4g1gOsi4krgFuDV\nFOePkiRJmlZGWXl6goh4BjDKqQp6FgbOzCvLx6k6tCRJUueq7vM0G1gJOLbKIBMKA0uSJE1bVfd5\nWgL8d2YuHXWAiYWBK8YnSZI0pYySPB2XmU9YaYqIT0y8rZc+hYElSZKmrVFOdvHCHrftOqzTgMLA\nkiRJ01bflafylAKvAzaLiAvH3bUmMMrZ6MYKA98cETeUt70fWAX4JLA+cFlE3JCZL51M8JIkSW0b\ntNnuNuAyYIfy/zEPAt8a9sBDCgNfMmqAkiRJU8nQ8iwRsU5m/r6leHqyPIs0dcyUUg0/WVT9+JXn\nzHtaA5E80WTKs0ymBM5UnZeZ8v7SzNCvPMso+zydERHrjF2JiHUnbMaTJEl60hgleXrG+JWnzPwd\n8KzmQpIkSZq6Rkme5kbEnLEr5ekHVmkuJEmSpKlrlPM8fR24ICI+UV4/DrhiWKcBhYFPAV4JPAb8\nAjgsM++fTPCSJEltG2Xl6f3AzcCp5b8bgK+O0G+sMPDWwI7A0RHxbOBKYJvM3JbiiL4TJhO4JElS\nF4auPGXm48CHIuIM4BDgMOBvgS2G9OtXGPib45r9AHjtJGOXJElq3cDkKSLmAvsAhwP/s2z/0sz8\nQZVBBhQGPhy4oMpjSZIkdanvZruIOBX4DfAW4IvAPOD3k0icehYGjogTKTbtnTeJuCVJkjoxaOXp\nrcB/Af+Ymd8GiIhKZy/rVxg4Ig4B9gb2yExPgClJkqaNQcnTxsBBwMciYm2KI+dGOToP6F8YOCL2\nAo4Hds3MUWrkSZIkTRlDy7MARMR8iv2T3gAspFhJ+j9D+vwN8D2KI/XG6g28H/hnivNE/a687QeZ\n+ZZBj2V5FmnqmCnlMyzPMjXnZaa8vzQz9CvPMlLyNKbcDLcfcGhmvrym2IYyeZKmjpny42byNDXn\nZaa8vzQz1JI8dcXkSXpymUwCMXfOKKetWzE33lH9fL7z/3qtBiKR1IYVKQwsSZKkksmTJElSBSMf\nPVfVgNp2fw/sW952D8X+U3c1FYckSVKdmlx56lfb7pTM3DYzFwCXAic1GIMkSVKtGkueMnNxZl5f\nXn6I4hQHm4w/yzjwFHBncEmSNH00ttluvIm17SLiwxTFhR8AdmsjBkmSpDo0vsN4r9p2mXliZm5K\nUdfumKZjkCRJqkujyVO/2nbjfBF4TZMxSJIk1amx5GlAbbstxjXbB7i1qRgkSZLq1uQ+TzsDBwM3\nR8QN5W3vB46IiKA4VcEdwMC6dpIkSVOJ5VkkTTmWZ5E0FVieRZIkqQauPEmq5NHHl1Zqv8pKcxqK\nZHp44OHHK7Vfc/WVKo+xdFn1r8g5s3v+Qa0ZbjK/+bNmPXnfK648SZIk1cDkSZIkqYLWCwOPu//d\nwCnA+pl5b1NxSJIk1amLwsBjidWLgV83OL4kSVLtWi8MXN59GvBeLAosSZKmmVb2eRpfGDgi9gHu\nzMwb2xhbkiSpTk2eYRx4YmFgik15JwIvaXpcSZKkJrRdGPiZwObAjRFxOzAPuD4iNmoyDkmSpLo0\ndpLMsjDwucDvM/PYPm1uB7YfdrSdJ8mUpg5PklmNJ8nUVOJJMqvpd5LM1gsDZ+blDY4pSZLUKMuz\nSKrEladqXHnSVOLKUzX9Vp5MnjSjLJvEj8hsf0QqqfpDPZlXd6rOyZKlyyr3mTun2q6lb/z8dZXH\n+MLfPr9yn6mq6m/SZH7Cpur7S1OPte0kSZJqYPIkSZJUgcmTJElSBa0XBo6IDwJHAb8tm3oEniRJ\nmjaaPFXBWGHg6yNiDeC6iLiyvO+0zPxYg2NLkiQ1orHkKTMXA4vLyw9FxPjCwJIkSdNS64WBy5uO\niYibIuLsiFi7jRgkSZLq0HjyNL4wcGY+CHyaosbdAoqVqY83HYMkSVJdmtznqVdhYDLz7nH3nwFc\n2mQMkiRJdWps5aksDHwWsDAzTx13+8bjmu0H3NJUDJIkSXVrvTAwcGBELACWA7cDb24wBkmSpFo1\nebTd9+ld1spzOkmSpGnLwsCaUawY3ryqr7Gvb/NuvOP+yn22ffqaldo7j3oysjCwJElSDUyeJEmS\nKmi9tl1539uBYyhKuFyWme9tKg5JkqQ6dVHbbkNgX2DbzHw0IjZoMAZJkqRatbbDeER8FfgX4CiK\nVaj/b9S+7jCuUbnDePPcYXzqcYdxqRmd7jA+obbdlsAuEXFNRHw3Il7QRgySJEl16KK23VxgbWBH\n4D3AheXZyCVJkqa8RpOnXrXtgEXAxZm5PDOvpdiZfL0m45AkSapL67XtgH8Hdi/bbAmsDNzbVByS\nJEl16qK23dnA2RFxC/AYcEhmukO4JEmaFizPohnFo+2a59F2U49H20nNsDyLJElSDVx50oyyZOmy\nyn3mzvEdQeyQAAAetUlEQVRviCoeX1LtNV5prq/vVHT/Hx+r1H6tp6zcUCTS1OXKkyRJUg1MniRJ\nkipovTBwRFwARNlsLeD+zFzQVBySJEl1ar0wcGa+fqxBRHwceKDBGCRJkmrVWPKUmYuBxeXlhyJi\nIbAJ8FP480k0X0d5wkxJkqTpoIvCwGN2Ae7OzJ+1EYMkSVIduigMPOZA4Pymx5ckSapTk/s89SsM\nTETMBV4NPL/J8SVJkurWRWFggD2BWzNzUVPjS5IkNaH1wsCZeTlwAG6ykyRJ05DlWTSjWJ6leZZn\nmRkszyIN1688i8mTJGmo6391X+U+222+dgORSO2xtp0kSVINTJ4kSZIq6KK23QLgM8CqFCVc3paZ\n1zYVhyRJUp2aXHkaq223NbAjcHREPBv4J+BDZTHgk8rrkiRJ00IXte2WA08rm60J3NVUDJIkSXVr\n9AzjYybUtjsW+EZEfIxi5WunNmKQJEmqQxe17d4KHJeZmwLHUZyFXJIkaVpoNHnqU9vuEGDs8peB\nHZqMQZIkqU5d1La7C9i1vLw78LOmYpAkSapb67XtgKOA0yNiLvAI8KYGY5AkSaqV5VkkSUNZnkVP\nRpZnkSRJqoErT5KkRiy+/5HKfTZac5VK7WfN6rkwINXClSdJkqQamDxJkiRV0EVh4PkUhYGfCtwO\nHFSePFOSJGnK66Iw8JnA+zLzucAlwHsajEGSJKlWjSVPmbk4M68vLz8EjBUGDuCqstmVwGuaikGS\nJKlurezzNKEw8C3APuVd+wObthGDJElSHbooDHw4xSa864A1gMeajkGSJKkuTZZn6VkYODNvBV5S\n3r8l8IomY5AkSapT64WBI2KD8v/ZwAcojryTJEmaFrooDLxFRBxdXr8YOKfBGCRJkmpleRZJUiMs\nz6LpzvIskiRJNXDlSZI0ZVz/q/sqtd9u87UbikRy5UmSJKkWJk+SJEkVNFkYeFWKMiyrlONclJkn\nR8TmwJeAdYDrgYMz0xNlSpKkaaHJladHgd0zcz6wANgrInYEPgqclplbAPcBRzQYgyRJUq2aLAy8\nPDP/UF5dqfy3HNgduKi8/VzgVU3FIEmSVLemy7PMAa4DngV8CvgFcH9mLimbLAI2aTIGSZKkOjW6\nw3hmLs3MBcA8YAdg6x7NPA2BJEmaNlo52i4z7we+A+wIrBURYyte84C72ohBkiSpDk0WBl4/ItYq\nL68G7AksBL4NvLZsdgjw1aZikCRJqluTK08bA9+OiJuAHwJXZualwPHAOyPi58C6wFkNxiBJklQr\ny7NIkqYMy7NoKulXnsXkSZI0ZVT9TfrSDb+pPMaBz3t65T56crK2nSRJUg1MniRJkiroorbdMcCx\nwDOB9TPz3qZikCRJqlsXte3+k+K0BXc0OLYkSVIjGlt5yszlwF/UtsvMHwNERFNDS5IkNabV2naZ\neU2T40mSJDWt1dp2EbFNk+NJkiQ1re3adnu1MZ4kSVJT2q5td2tT40mSJLWh9dp2EfF3EbGIYlPe\nTRFxZoMxSJIk1cryLJKkKcPyLJpKLM8iSZJUA1eeJElPKovvf6RS+43XWrWhSLQiquYvs2b1XEQa\nyJUnSZKkGpg8SZIkVdBFYeDzgO2Bx4FrgTdn5uNNxSFJklSnLgoDnwdsBTwXWA04ssEYJEmSatVF\nYeDLx9pExLUU53uSJEmaFjorDBwRKwEHA+9oMgZJkqQ6dVkY+F+BqzLze03GIEmSVKdOCgNHxMnA\n+sA72xhfkiSpLq0XBo6II4GXAgdm5rKmxpckSWpC64WBgc8AGwJXR8QNEXFSgzFIkiTVyvIskqQn\nFcuzzAxdlmcxeWrZsmXVnsrs2dUnW1PPZD5nk/mgS6rfJTctqtxnv22rn4WnjWSgje+iqr9zMHV/\n66xtJ0mSVAOTJ0mSpApMniRJkiroojDwWRSFgWcBtwGHZuYf+j+SJEnS1NFFYeDjMnN+Zm4L/Bo4\npsEYJEmSatVFYeAHASJiFrAazJwj6SRJ0szXSWHgiDgHeDnwU+BdTcYgSZJUp04KA2fmYcBfAQuB\n1zcZgyRJUp06KQxc3rYUuAB4TRsxSJIk1aHtwsAZEc8qb5sFvBK4takYJEmS6tbkPk8bA+eW+z3N\nBi4ELgO+FxFPozhVwY3AWxuMQZIkqVZNHm13E/C8Hnft3NSYkiRJTbMwsCRJNfvpogcr93n2vKc1\nEMmKe+TxpZXar7rSnIYiaZ+FgSVJkmpg8iRJklRB67Xtxt3/SeCwzHxqUzFIkiTVrYvadkTE9sBa\nDY4tSZLUiNZr25WnLjgFeAOwX1PjS5IkNaHRfZ4iYk5E3ADcA1xZ1rY7BvhaZi5ucmxJkqQmNFoY\nuCzBsqA80/glEfFCYH/gRU2OK0mS1JS2a9vtBjwL+HlE3A6sHhE/byMGSZKkOjR5tN36wOOZef+4\n2nYfzcyNxrX5Q2Y+q6kYJEmS6tZqbbvMvLTB8SRJkhpneRZJkmpmeZaZwfIskiRJNXDlSZKkKeCd\nX/1ppfan7vvshiLRGFeeJEmSamDyJEmSVEHrhYEj4nPArsADZdNDM/OGpuKQJEmqU5OnKhgrDPyH\niFgJ+H5EXFHe957MvKjBsSVJkhrRemHgpsaTJElqQxeFgQE+HBE3RcRpEbFKkzFIkiTVqdHkKTOX\nZuYCYB6wQ0RsA5wAbAW8AFgHOL7JGCRJkurUdmHgvTJzcWYuz8xHgXOAHdqIQZIkqQ6NJU8RsX5E\nrFVeHisMfGtEbFzeNgt4FXBLUzFIkiTVrfXCwBHxHxGxPjALuAF4S4MxSJIk1arJo+1uAp7X4/bd\nmxpTkiSpada2kyRpGrr+V/dV7rPd5ms3EMnMZW07SZKkGpg8SZIkVWDyJEmSVEEXhYFnAf8A7A8s\nBT6dmf/cVBySJEl1anLlaaww8HxgAbBXROwIHApsCmyVmVsDX2owBkmSpFp1URj4rcAbMnNZ2e6e\npmKQJEmqWxeFgZ8JvD4ifhQRV0TEFk3GIEmSVKcuCgOvAjySmdsDZwBnNxmDJElSnVovDAwsAr5S\n3nUJsG0bMUiSJNWh9cLAwL8DYyVadgVuayoGSZKkunVRGPj7wHkRcRzFDuVHNhiDJElSraxtJ0nS\nNGRtu+b1q21n8iRVNJnPzKxZPT9/ktSqxfc/UrnPxmut2kAk04OFgSVJkmpg8iRJklRBF7Xtvges\nUTbbALg2M1/VVBySJEl1avJou7Hadn+IiJWA70fEFZm5y1iDiPgK8NUGY5AkSapVF7XtAIiINSjO\n93RYUzFIkiTVrcmVJ8pzPF0HPAv4VFnbbsx+wLcy88EmY5AkSapTF7XtxhwInN/k+JIkSXXrorYd\nEbEusANwWRvjS5Ik1aWL2nYA+wOXZmb1s3VJkiR1qPXaduV9BwAfaXBsSZKkRlieRarI8iySpivL\ns1RjeRZJkqQauPIkSVLNZtIK9dr7n1Gp/X1fPqqhSNrnypMkSVINTJ4kSZIq6KIw8B7AKRSJ2x+A\nQzPz503FIUmSVKcmV57GCgPPBxYAe0XEjsCngYPKM49/EfhAgzFIkiTVqovCwMuBp5W3rwnc1VQM\nkiRJdWu9MHBEHAlcHhF/Ah4EdmwyBkmSpDp1URj4OODlmTkPOAc4tckYJEmS6tR2YeCXAfMz85ry\nrguAndqIQZIkqQ5tFwZeCKwZEVuWzV5c3iZJkjQttF4YOCKOAr4SEcuA+4DDG4xBkiSpVpZnkSSp\nZpZnmRn6lWcxeZIkSbW5674/Ve7zV2uv1kAkK87adpIkSTUweZIkSarA5EmSJKmCLgoD7w58DFiZ\n4uzjR2TmkqbikCRJqlPbhYF3As4FDsjMbYA7gEMajEGSJKlWjSVPmbk8MycWBl4KPJqZt5W3Xwm8\npqkYJEmS6tboPk8RMScibgDuoUiUrgVWiojtyyavBTZtMgZJkqQ6tVoYGHgOcABwWkRcCzwEuL+T\nJEmaNtouDLxXZl6dmbtk5g4UO5T/rI0YJEmS6tB2YeBbI2KD8rZVgOOBzzQVgyRJUt2aXHnaGPh2\nRNwE/BC4MjMvBd4TEQuBm4D/m5n/0WAMkiRJtbK2nSRJqs2TobbdjE2eqj6vqVrNui1P5tfrT48t\nrdR+tZXnNBSJJE09y5ZV+32YPbv678Ntix+q3GfLjdeo3KcqCwNLkiTVwORJkiSpgsZq242JiDnA\nj4A7M3PviNgc+BKwDnA9cHBmPtZ0HJIkSXVoY+XpHcDCcdc/CpyWmVsA9wFHtBCDJElSLZouzzIP\neAVwZnl9FrA7cFHZ5FzgVU3GIEmSVKemV54+AbwXWFZeXxe4PzPHSrIsAjZpOAZJkqTaNHmG8b2B\nezLzunE39zrkb+qfK0GSJKnU5MrTzsA+EXE7xQ7iu1OsRK0VEWM7qs8D7mowBkmSpFo1ljxl5gmZ\nOS8zNwMOAP4jMw8Cvg28tmx2CPDVpmKQJEmqWxfneToeeGdE/JxiH6izOohBkiRpUizPUppJ5UYm\n48n8elmeRZL6szzLX/IM45IkSRXM2JUnSZJ6eXzJsuGNxllprusMU9G9Dz1aqf16a6xSeQxXniRJ\nkmpg8iRJklRBF4WBjwGOBZ4JrJ+Z9zYdgyRJUl26KAz8n8CewB0tjC1JklSrVgsDA2TmjzPz9ibH\nlSRJakrbhYElSZKmtbYLA0uSJE1rrRYGjogvNDieJElS4xo72i4zTwBOAIiIFwHvzsw3NjWeJElS\nG1o/z1NE/F1ELALmATdFxJnD+kiSJE0VlmeRJD2pWJ5lZuiyPIvJkyRJmvHuuu9Plfs8Y/3VrG0n\nSZK0okyeJEmSKuiitt15wPbA48C1wJsz8/Gm45AkSapDF7XtzgO2Ap4LrAYc2UIMkiRJteiitt3l\nmbk8M5dTrDzNazIGSZKkOnVW2y4iVgIOBr7ecAySJEm16bK23b8CV2Xm95qKQZIkqW6d1LaLiJOB\n9YF3Nji+JElS7Vo5Sea42nZ7R8SRwOHAHpk50hmrPEmmJElaEdP9JJmfATYEro6IGyLipA5ikCRJ\nmhTLs0iSpBlvuq88SZIkTVvTYuVJkiRpqnDlSZIkqQKTJ0mSpApMniRJkioweZIkSarA5EmSJKkC\nkydJkqQKTJ4kSZIqmNt1AJMVEXsBpwNzgDMz8yND2p8N7A3ck5nbjPD4mwKfBzYClgGfzczTh/RZ\nFbgKWIXitb0oM08eYaw5wI+AOzNz7xHa3w48BCwFlmTm9iP0WQs4E9gGWA4cnplX92kbwAXjbnoG\ncFJmfmLIGMcBR5aPfzNwWGY+MqD9O4CjgFnAGf0ev9fcRcQ6ZYybAbcDr8vM+wa03x/4ILA1sENm\n/miEMU4BXgk8BvyifD73D2j/98C+FO+Xe4BDM/OuQWOMu+/dwCnA+pl575C4Pli+br8tm70/My8f\nNEZEvB04BlgCXJaZ7x0yxgVAlE3WAu7PzAVD+iygKL+0ajnO2zLz2gHt55ftn0oxhwdl5oPjxuj5\nGew39wPa9537AX16zv2A9n3nfth3Sa+5HzDOB+kx94PG6Df3A8boO/cD+vSc+wHt+859v+/RiNic\nosj8OsD1wMGZ+diQPscAxwLPnPD69mt/HrA98DhwLfDmzHx8SJ+zyj6zgNvKuf/DoD7j5v6TFO+t\npw4Z43PArsADZddDM/OGIX1mAf8A7E/xW/HpzPznAe2/B6xRPv4GwLWZ+aohY+xB8d6dDfyhjOvn\nQ/rsDnwMWBm4DjgiM5eMe02e8Hs4aN4H9Ok57wPa9533fqblylP5xD8FvAx4NnBgRDx7SLfPAXtV\nGGYJ8K7M3BrYETh6hDEeBXbPzPnAAmCviNhxhLHeASysEBvAbpm5YJTEqXQ68PXM3AqYP2i8LCwo\nvzCfDzwMXDLowSNiE+DvgO3LH8c5wAED2m9D8SOwQxnP3hGxRZ/mn+Mv5+59wLcycwvgW+X1Qe1v\nAV5N8WEedYwrgW0yc1uKL8UThrQ/JTO3LV+3S4GJdRt79Rn7QXox8OsR4wI4bWyOxhKnfu0jYjeK\nH/ZtM/M5FF9cA8fIzNePew98Bbh4hLj+CfhQ2eek8vqg9mcC78vM51K8v94z4f5+n8F+c9+v/aC5\n79en39z3az9o7vt+lwyY+0HfP73mvmf7IXPfs8+Que8XV7+579d+0Nz3+x79aPnctwDuA44Yoc9/\nAnsCd0x4ffu1Pw/YCngusBrFH4PD+hyXmfPL98qvKRLVYX2IiO0pktNR4gJ4z7h5v2GEPocCmwJb\nla//lwa1z8xdxs371Txx3vuN8WmKxHcB8EXgA0P67AScCxxQ/k7cARwy4TWY+Hs4aN779ek37/3a\nD5r3nqZl8kTxg/vzzPxlmYF+ieILoq/MvAr4/agDZObizLy+vPwQxQu9yZA+y8f+4gBWKv8NPIV7\nRMwDXkHxZdKIiHga8ELgrDLOx7JcQRnBHsAvMrPfm3C8ucBqETEXWB24a0DbrYEfZObD5V8d3wX2\n69Wwz9ztS/EhpPz/VYPaZ+bCzMx+wfTp881xfxH9AJg3pP2D464+hQlzP+A9eBrw3onth/QZ+XkA\nbwU+kpmPlm3uGXWM8q/X1wHnj9BnOfC08vKajJv/Pu2D/5fQXAm8ZsIY/T6DPee+X/tBcz+gT8+5\nH9C+79wP+S7pOfdVv38GtO8798PG6DX3A/r0nPsB7fvO/YDv0d2Bi8rbJ37me/bJzB9n5u09Xq9+\n7S8v71tOsQIxb4Q+Yytmsyh+eJcP61MuAJxCMfdD45oY/4h93gr8r8xcVra7Z5QxImINitf630cY\nY9BnvlefpcCjmXlbefsT5n7i72H5mvad9159yrF7zvuA9n3nvZ/pmjxtAvxm3PVFDElsVkREbAY8\nD7hmhLZzIuIGiqX7KzNzWJ9PUHyAllUIaTnwzYi4LiLeNEL7Z1As858TET+OiDMj4ikjjnUAE344\ne8nMOyn+qv01sBh4IDO/OaDLLcALI2LdiFgdeDnFX0mj2jAzF5djL6ZYZm7S4cAVwxpFxIcj4jfA\nQfzlylOv9vtQLB3fWDGeYyLipog4OyLWHtJ2S2CXiLgmIr4bES+oMM4uwN2Z+bMR2h4LnFI+/4/x\nxJW6Xm4B9ikv78+A+Z/wGRw691U+syP06Tn3E9uPMvfj+4w69z3iGjj3E9qPNPd9nvvAuZ/QZ+jc\nT2g/cO4nfo9SbDq9f1xC+xff+VW/ewe1j4iVgIOBr4/SJyLOAf6bYvXikyP0OQb42tj7eMS4PlzO\n+2kRscoIfZ4JvD4ifhQRV4xf2R/yWu1HsbL74AhjHAlcHhGLytfrI4P6UCQmK5WrbgCv5YlzP/H3\ncF2GzHuPPsP0bd9v3nuZrslTryrHjRTpi4inUixdHzvxzdRLZi7NYglzHrBDuXmq32OP7QNyXcWw\nds7M7Sg2Wx4dES8c0n4usB3FNu/nAX/kiZu5+sW3MsUX3JdHaLs2xYrA5sBfAU+JiDf2a5+ZCymW\nY6+keKPeSLHEP+VExIkUsZ03rG1mnpiZm5ZtjxnUtkwaT2SEJGuCT1N8MS6gSFQ/PqT9XGBtis0m\n7wEuLP+iG8WBjJA8l95KsQljU+A4ypXOAQ6neP9eR7GvxWO9GlX9DFZtP6hPv7nv1X7Y3I/vUz7m\n0LnvMc7Aue/RfujcD3i9+s59jz4D575H+4FzP/F7lGKleqKJq3Ujf/eO0P5fgasy83uj9MnMwyi+\n9xYCrx/S54UUCeMTkqwhY5xAkZi9gGLfn+NH6LMK8EgWu3acAZw94nPvOe99+hwHvDwz5wHnAKcO\n6gM8h+IP8tMi4lqKfXeXQN/fw4G/9VV/Q0do33Pee5muydMinpitzmPwJqJJKbPQrwDnZebEfT4G\nymKz2HcYvJ/VzsA+UewA/iVg94j4wgiPPbYkfg/F/gI7DOmyCFg07q+LiyiSqWFeBlyfmXeP0HZP\n4FeZ+dssdrS7GNhpUIfMPCszt8vMF1Js0hlldWPM3RGxMUD5/z1D2k9KRBxCsaPzQeWS7qi+yITN\nUD08kyLZvLF8D8wDro+IjQZ1ysy7yy+lZRRfiqPM/8XlsvS1FH9xrTfsCZSbX1/NEw8eGOQQ/t9+\nEl8eFldm3pqZL8nM51N8Wf+iRwy9PoN9534yn9l+ffrN/Qhj/MXc9+gzdO57jTNo7vvENXDuBzz3\nvnPfp0/fue/zPIbOfdlu7Ht0R2CtMi4Y8J0/4ndv3/YRcTKwPvDOUfuUty2leL16fu7H9dkNeBbw\n83LuV4+Inw8aI4vNn8uz2Px6Dn0+WxPiWkTxukPxO7HtCM993fKxLxvhub8MmD/ud+UC+nznT3gu\nV2exf9UOFJtux773/+L3kGKVaNC8V/0N7dt+lHkfb7omTz8EtoiIzcvVkQOAr9U5QPnX2VnAwsw8\ndVj7ss/6URzVRkSsRpFQ3NqvfWaekJnzMnMziufwH5nZd7WmfNynRLFNmig2vb2EYgm8r8z8b+A3\nETF2BM0ewE9HeEpVVh1+DewYEauXr90eDNkJPiI2KP9/OsUX9ahjQTHfYzsaHgJ8tULfkURxROfx\nwD6Z+fAI7cfv8L4PA+YeIDNvzswNMnOz8j2wCNiunK9B42w87up+DJl/in0Xdi/7bklxlMu9A3sU\n9gRuzcxFI7SF4ktt1/Ly7gxJhsfN/2yKHU0/M+H+fp/BnnM/yc9szz795n5A+75z36vPsLkfME7P\nuR/w3PvO/ZDXq+fcD+jTc+4HPI++c9/ne3Qh8G2KzTww4TNf9bu3X/uIOBJ4KXBgmaAO65MR8axx\nz/WVPHHue/W5LjM3Gjf3D2fmswa0v3XcHwuzKPb5uWXIGLcybu4p5ua2EV6r/YFLc8JR0gPmZM3y\nfQXFgQ8Lh/S5ddzcr0LxGfsM9P09PIgB8171N7Rf+0Hz3s+0PFVBZi6J4lDEb1Ac1XV2Zv5kUJ+I\nOB94EbBeFNtnT87MQZsVdqbY9nlzFNtsYdwh4X1sDJwbxc6As4ELM/PSkZ7U6DYELinzoLnAFzNz\n6PZZ4O3AeWWy+UvgsEGNo9ik9GLgzaMElZnXRMRFFIeSLgF+/P+3d38hVpRhHMe/q0VFRgUtCFbQ\nhfzIGy8iCExJKCTpj9AmaWqWWtpS4F0RdFHSTRD0D0JI6A+4SXlhtDfiYir9waBNl+K5CKKLQJAg\n7WLFcLt43pXjeubMzrKye9bf53LmnXln5t0dnvPOzPMAu2s2+6r80jkP9EdJNdDmWC4bO/LZ+j5J\nW8jA7cma9n+T0+S9wDeShiNiVc02r5JT3wfL9f4hIrZ3aL+6BKgXyK88ttedR83fYFU/Dyg/DR8j\nP/N+oab9HmCPpBHy8cgzE2ZSqo6r8n23in62Ae+WX4mjwPM17RdI6i9N9pO/qFu1/R+keuyr2l9H\n9dhXbfMe7ce+qv2WDmM/lXtJVT/rKsa+qn2nse90XFVjX9VP1dhXtV/cYezb3kcl/QoMSNpF3l8+\nnsQ2L5PvtywETkgajIitHdr/R47f92Xc90fEG1V9kDM0R5Uf5fSQrx/sqDuuNte17jyGJPWWPoa5\n9O+raptj5D1/J5lGYGun9mXdU0x4b6mmj23kffwC+SXcc5PY5m3l47N55KskQx2uB2SAVTXubXUY\n9yofUT3ubfWMjV2RV4XMzMzM5qRufWxnZmZmNiMcPJmZmZk14ODJzMzMrAEHT2ZmZmYNOHgyMzMz\na6ArUxWY2dygTFY3ShYRnQ/sioiBTttMcp+PRMSIpEHgpYhom4SxtF8D/FWSSDbta3Ppq6+urZnN\nHZ55MrOZ1hdZeX0jWX/xkuznJU/MlETE6k6BU7GG+iztZmYXeebJzGaFiPhZ0lngWUmryLpXi4EN\nkk6RiS7vJCvX742ItwAkLSdrUgF8S0s9rAmzUIvI5Jfj2cD3kkldHwMeLFmG34mIT5WlWV4k75H/\nADsiIkqS2ffJzM2nyaR9ZnaVcfBkZrOCpJXA9WTG+fvIulm/l3UHgTcj4kgJYA5JOk7Wxhog688d\nlrQW6G/fA58DgxHxRNnnbRFxWtIB4KeI+KAsXw6sBVZExDlJD5OZupeRGb3vApYA15b+/5jua2Fm\ns5uDJzObaV9KGgXOkEVVFwHHWgKnG8nSLr26WJ6Rm4C7gVNkbbDDABGxT9JlZYEkLSCLlj40viwi\nqur7PQosBX4s/fUAt5Z1K4FPIotfn1cWFb1/SmdtZl3LwZOZzbS+iGgtdLqZrMU1bh5Zy+3eErTQ\n0nbpFTieHrJe5usV68zsKucXxs1sVouIs8BR4JXxZZLukLSQrAZ/g6QVZXkfcHObffwLfAfsbNnH\n+IvpZyZs8zWwSdLtpd18SfeUdYeAjZKuUVaKXz89Z2lm3cTBk5l1g6eBJZJOSjoJfAHcEhHngHXA\nh5JOkI/3/qzYxwZgmaQRSb8AW8ryz4D1koYlbYqII8BrwIHSbgR4vLTdXfb/GzAEHJ/uEzWz2a9n\nbGxspo/BzMzMrGt45snMzMysAQdPZmZmZg04eDIzMzNrwMGTmZmZWQMOnszMzMwacPBkZmZm1oCD\nJzMzM7MG/geb8EXcVIpEPgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f68c82e9e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.xticks(range(43))\n",
    "plt.yticks(range(43))\n",
    "plt.grid(False)\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.title('Confusion Matrix')\n",
    "plt.imshow(confusion_matrix(y_valid, validation_pred_label[0]), cmap=plt.cm.Blues)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7f68a20f5278>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/carnd/anaconda3/envs/carnd-term1/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nh5mZFeLEYWZmhfwfR7QTSO1lO4UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f68c9d26d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(range(1, EPOCHS+1), training_acc_hist, label = \"Training\")\n",
    "plt.plot(range(1, EPOCHS+1), validation_acc_hist, label = 'Validation')\n",
    "plt.title(\"Training and Validation Accuracy\")\n",
    "plt.legend()\n",
    "plt.xlabel(\"Epoch\")\n",
    "plt.ylabel(\"Accuracy (%)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test set accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy = 0.935\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "\n",
    "    test_accuracy = evaluate(X_test, y_test)\n",
    "    print(\"Test Accuracy = {:.3f}\".format(test_accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "---\n",
    "\n",
    "## Step 3: Test a Model on New Images\n",
    "\n",
    "To give yourself more insight into how your model is working, download at least five pictures of German traffic signs from the web and use your model to predict the traffic sign type.\n",
    "\n",
    "You may find `signnames.csv` useful as it contains mappings from the class id (integer) to the actual sign name."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Load and Output the Images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from mpl_toolkits.axes_grid1 import ImageGrid\n",
    "\n",
    "## read in the new images\n",
    "def read_new_images(files_path):\n",
    "    images = []\n",
    "    \n",
    "    for file_path in files_path:\n",
    "        img = cv2.cvtColor(cv2.imread(file_path),cv2.COLOR_BGR2RGB)\n",
    "        img = cv2.resize(img, (image_shape[0], image_shape[1]))\n",
    "        images.append(img)\n",
    "        \n",
    "    return np.array(images)\n",
    "\n",
    "def plot_img_grid(img_array):\n",
    "\n",
    "    fig = plt.figure(figsize = (10,15))\n",
    "    \n",
    "    grid = ImageGrid(fig, 111, nrows_ncols=(2, 3), axes_pad=0.3,)\n",
    "\n",
    "    for i in range(len(img_array)):\n",
    "        grid[i].grid(False)\n",
    "        grid[i].set_title(str(sign_names[sign_names[\"ClassId\"] == new_labels[i]].iloc[0,1]) + \" - Class \" + str(new_labels[i]))\n",
    "        grid[i].imshow(img_array[i])\n",
    "\n",
    "    plt.show()\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/carnd/anaconda3/envs/carnd-term1/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAloAAAGcCAYAAAABAg7eAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f68c9f722e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Load and plot the new test images\n",
    "sign_names = pd.read_csv('/home/carnd/sdcnd_traffic_sign/signnames.csv') ## on server\n",
    "new_images_path = ['./media/new_image_1.png','./media/new_image_2.png','./media/new_image_3.png','./media/new_image_4.png',\n",
    "         './media/new_image_5.png','./media/new_image_6.png']\n",
    "new_labels = [25, 17, 5, 24, 18, 12]\n",
    "\n",
    "new_images = read_new_images(new_images_path)\n",
    "plot_img_grid(new_images)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Predict the Sign Type for Each Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "New Image Accuracy = 0.833\n"
     ]
    }
   ],
   "source": [
    "### Run the predictions here and use the model to output the prediction for each image.\n",
    "### Make sure to pre-process the images with the same pre-processing pipeline used earlier.\n",
    "### Feel free to use as many code cells as needed.\n",
    "new_images_processed = preprocess(new_images)[:,:,:,np.newaxis]\n",
    "\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "\n",
    "    new_img_accuracy = evaluate(new_images_processed, new_labels)\n",
    "    print(\"New Image Accuracy = {:.3f}\".format(new_img_accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Output Top 5 Softmax Probabilities For Each Image Found on the Web"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "For each of the new images, print out the model's softmax probabilities to show the **certainty** of the model's predictions (limit the output to the top 5 probabilities for each image). [`tf.nn.top_k`](https://www.tensorflow.org/versions/r0.12/api_docs/python/nn.html#top_k) could prove helpful here. \n",
    "\n",
    "The example below demonstrates how tf.nn.top_k can be used to find the top k predictions for each image.\n",
    "\n",
    "`tf.nn.top_k` will return the values and indices (class ids) of the top k predictions. So if k=3, for each sign, it'll return the 3 largest probabilities (out of a possible 43) and the correspoding class ids.\n",
    "\n",
    "Take this numpy array as an example. The values in the array represent predictions. The array contains softmax probabilities for five candidate images with six possible classes. `tk.nn.top_k` is used to choose the three classes with the highest probability:\n",
    "\n",
    "```\n",
    "# (5, 6) array\n",
    "a = np.array([[ 0.24879643,  0.07032244,  0.12641572,  0.34763842,  0.07893497,\n",
    "         0.12789202],\n",
    "       [ 0.28086119,  0.27569815,  0.08594638,  0.0178669 ,  0.18063401,\n",
    "         0.15899337],\n",
    "       [ 0.26076848,  0.23664738,  0.08020603,  0.07001922,  0.1134371 ,\n",
    "         0.23892179],\n",
    "       [ 0.11943333,  0.29198961,  0.02605103,  0.26234032,  0.1351348 ,\n",
    "         0.16505091],\n",
    "       [ 0.09561176,  0.34396535,  0.0643941 ,  0.16240774,  0.24206137,\n",
    "         0.09155967]])\n",
    "```\n",
    "\n",
    "Running it through `sess.run(tf.nn.top_k(tf.constant(a), k=3))` produces:\n",
    "\n",
    "```\n",
    "TopKV2(values=array([[ 0.34763842,  0.24879643,  0.12789202],\n",
    "       [ 0.28086119,  0.27569815,  0.18063401],\n",
    "       [ 0.26076848,  0.23892179,  0.23664738],\n",
    "       [ 0.29198961,  0.26234032,  0.16505091],\n",
    "       [ 0.34396535,  0.24206137,  0.16240774]]), indices=array([[3, 0, 5],\n",
    "       [0, 1, 4],\n",
    "       [0, 5, 1],\n",
    "       [1, 3, 5],\n",
    "       [1, 4, 3]], dtype=int32))\n",
    "```\n",
    "\n",
    "Looking just at the first row we get `[ 0.34763842,  0.24879643,  0.12789202]`, you can confirm these are the 3 largest probabilities in `a`. You'll also notice `[3, 0, 5]` are the corresponding indices."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy = 0.833\n"
     ]
    }
   ],
   "source": [
    "### Print out the top five softmax probabilities for the predictions on the German traffic sign images found on the web. \n",
    "### Feel free to use as many code cells as needed.\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "\n",
    "    logits = sess.run([tf.nn.softmax(logits)], feed_dict={x: new_images_processed, y: new_labels, keep_prob: 1.0})\n",
    "    print(\"Test Accuracy = {:.3f}\".format(new_img_accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f68a06ba128>"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/carnd/anaconda3/envs/carnd-term1/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAFgCAYAAACrJILeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIpIJBXQRkUwooIuIZEIBXUQkE/8fvz9pR1vlI54AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f68a1be7128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plt.bar(range(len(logits[0][0].argsort()[-5:][::-1])),logits[0][0][logits[0][0].argsort()[-5:][::-1]])\n",
    "# plt.xlabels()\n",
    "#sign_names[sign_names['ClassId'].isin(logits[0][0].argsort()[-5:][::-1])]\n",
    "ClassId = logits[0][0].argsort()[-5:][::-1]\n",
    "probs = logits[0][0][logits[0][0].argsort()[-5:][::-1]]\n",
    "top5_df = pd.DataFrame({\"ClassId\":ClassId, \"Probs\":probs})\n",
    "\n",
    "filtered_signs = sign_names[sign_names['ClassId'].isin(logits[0][0].argsort()[-5:][::-1])]\n",
    "\n",
    "merged_df = pd.merge(filtered_signs, top5_df)\n",
    "\n",
    "merged_df.plot(x=\"SignName\", y=\"Probs\", kind=\"bar\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def plot_softmax_probs(logits_array, img_list, img_labels):\n",
    "        \n",
    "    for i, img in enumerate(logits_array[0]):\n",
    "        plt.subplot(121)\n",
    "        plt.grid(False)\n",
    "        plt.title(str(sign_names[sign_names[\"ClassId\"] == img_labels[i]].iloc[0,1]) + \" - Class \" + str(img_labels[i]))\n",
    "        plt.tick_params(axis='x', colors=(0,0,0,0))\n",
    "        plt.tick_params(axis='y', colors=(0,0,0,0))\n",
    "        plt.imshow(img_list[i])\n",
    "        \n",
    "        plt.subplot(122)\n",
    "        ClassId = img.argsort()[-5:][::-1]\n",
    "        probs = img[img.argsort()[-5:][::-1]]\n",
    "        top5_df = pd.DataFrame({\"ClassId\":ClassId, \"Probs\":probs})\n",
    "\n",
    "        filtered_signs = sign_names[sign_names['ClassId'].isin(img.argsort()[-5:][::-1])]\n",
    "\n",
    "        merged_df = pd.merge(filtered_signs, top5_df)\n",
    "\n",
    "        plt.bar(range(len(merged_df['Probs'])), merged_df[\"Probs\"])\n",
    "        plt.xticks(range(len(merged_df['SignName'])), list(merged_df['SignName']), rotation = 90)\n",
    "        plt.show()\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/carnd/anaconda3/envs/carnd-term1/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAFxCAYAAABqRfOaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ns6gxBMpUVQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f689bf2acc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAE7CAYAAADEhtqQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRKSk/j+EH7p41ailegAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f68a0266d30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAFlCAYAAADyArMXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFBlwIYQoKDLgQghRUGTAhRCioPz/ZWWg5hklyacAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f689bfbf278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAFxCAYAAACbTqgXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4C8iUkAK/iIiBfT/AWzbaskGyfj1AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f689bc17208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAGmCAYAAABlft8NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nbsi3/QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f689bdfb9b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAHrCAYAAADWuMa/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqdFOC5s3KgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f68a0064e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_softmax_probs(logits, new_images, new_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "> **Note**: Once you have completed all of the code implementations and successfully answered each question above, you may finalize your work by exporting the iPython Notebook as an HTML document. You can do this by using the menu above and navigating to  \\n\",\n",
    "    \"**File -> Download as -> HTML (.html)**. Include the finished document along with this notebook as your submission."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Project Writeup\n",
    "\n",
    "Once you have completed the code implementation, document your results in a project writeup using this [template](https://github.com/udacity/CarND-Traffic-Sign-Classifier-Project/blob/master/writeup_template.md) as a guide. The writeup can be in a markdown or pdf file. "
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.5.2"
  },
  "widgets": {
   "state": {},
   "version": "1.1.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}