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": 3,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Load pickled data\n",
    "import pickle\n",
    "\n",
    "dirpath = \"data\"\n",
    "\n",
    "training_file = dirpath + \"/train.p\"\n",
    "validation_file= dirpath + \"/valid.p\"\n",
    "testing_file = dirpath + \"/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",
    "#Load human readable lables\n",
    "import csv\n",
    "labels_human = []\n",
    "with open('signnames.csv') as sfile:\n",
    "    reader = csv.DictReader(sfile)\n",
    "    for row in reader:\n",
    "        labels_human.append(row['SignName'])"
   ]
  },
  {
   "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": 4,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training examples = 34799\n",
      "Number of testing examples = 12630\n",
      "Image data shape = (32, 32, 3)\n",
      "Number of classes = 43\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "n_train = len(X_train)\n",
    "n_test = len(X_test)\n",
    "\n",
    "image_shape = np.shape(X_train[0])\n",
    "\n",
    "n_classes = len(np.unique(y_train))\n",
    "\n",
    "print(\"Number of training examples =\", n_train)\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": 5,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "27 = Pedestrians\n",
      "5 = Speed limit (80km/h)\n",
      "38 = Keep right\n",
      "2 = Speed limit (50km/h)\n",
      "25 = Road work\n",
      "10 = No passing for vehicles over 3.5 metric tons\n",
      "18 = General caution\n",
      "38 = Keep right\n",
      "35 = Ahead only\n",
      "4 = Speed limit (70km/h)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAADfCAYAAADmzyjKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKFGibBCJC3qUKFGibBD5f+6h/tkcCP/ZAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd2622ffd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd27c94668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd28e2cbe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd28e13c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from random import randint\n",
    "import matplotlib.pyplot as plt\n",
    "# Visualizations will be shown in the notebook.\n",
    "%matplotlib inline\n",
    "\n",
    "#Display 10 random samples and their labels from the training set\n",
    "fig, axes = plt.subplots(2,5)\n",
    "for i in range(0,10):\n",
    "    index = randint(0,n_train)\n",
    "    axes[int(i/5 ), i%5].set_title(\"Label: \" + str(y_train[index]))\n",
    "    axes[int(i/5 ), i%5].imshow(X_train[index])\n",
    "    print(y_train[index],\"=\", labels_human[y_train[index]])\n",
    "    i = i+1\n",
    "    \n",
    "#Chart of class distribution\n",
    "plt.show()\n",
    "plt.hist(y_train, normed=True, bins=range(0,43))\n",
    "plt.title(\"Distribution of labels in the training data set\")\n",
    "plt.show()\n",
    "\n",
    "plt.hist(y_valid, normed=True, bins=range(0,43))\n",
    "plt.title(\"Distribution of labels in the validation data set\")\n",
    "plt.show()\n",
    "\n",
    "plt.hist(y_test, normed=True, bins=range(0,43))\n",
    "plt.title(\"Distribution of labels in the test data set\")\n",
    "plt.show()"
   ]
  },
  {
   "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": 7,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1bd2953a6d8>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAACRCAYAAADTnUPWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd29205128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import cv2\n",
    "\n",
    "def grayscale(img):\n",
    "    return cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
    "\n",
    "def process(imgs, labels):\n",
    "    outp = []\n",
    "    labp = []\n",
    "    for inx in range(0,len(imgs)):\n",
    "        zeros = np.zeros((32,32,1))\n",
    "        norm = cv2.normalize(grayscale(imgs[inx]), zeros, alpha=0, beta=1, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_32F)\n",
    "        outp.append(norm)\n",
    "        labp.append(labels[inx])\n",
    "        #outp.append(cv2.flip(norm,1))\n",
    "        #labp.append(labels[inx])\n",
    "    return np.array(outp), np.array(labp)\n",
    "\n",
    "index = randint(0,n_train)\n",
    "fig, axes = plt.subplots(1,3)\n",
    "axes[0].set_title(\"Original\")\n",
    "axes[0].imshow(X_train[index])\n",
    "\n",
    "axes[1].set_title(\"Grayscale\")\n",
    "axes[1].imshow(grayscale(X_train[index]))\n",
    "\n",
    "zeros = np.zeros((32,32,1))\n",
    "axes[2].set_title(\"Normalized\")\n",
    "axes[2].imshow(cv2.normalize(grayscale(X_train[index]), zeros, alpha=0, beta=1, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_32F))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Model Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(?, 30, 30, 6)\n",
      "(?, 26, 26, 16)\n",
      "(?, 13, 13, 16)\n",
      "(?, 9, 9, 24)\n",
      "(?, 4, 4, 24)\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.contrib.layers import flatten\n",
    "\n",
    "def LeNet(x):\n",
    "    #Expose this layer to visualize later (See Step 4)\n",
    "    global conv_layer2\n",
    "    \n",
    "    mu = 0\n",
    "    sigma = 0.1\n",
    "    \n",
    "    keep_prob = 0.6\n",
    "\n",
    "    filter_size_height = 3\n",
    "    filter_size_width = 3\n",
    "    k_output = 6\n",
    "    color_channels = 1\n",
    "\n",
    "    weight1 = tf.Variable(tf.truncated_normal(\n",
    "    [filter_size_height, filter_size_width, color_channels, k_output], mean=0, stddev=0.1))\n",
    "    bias1 = tf.Variable(tf.zeros(k_output))\n",
    "\n",
    "    conv_layer1 = tf.nn.conv2d(x, weight1, strides=[1, 1, 1, 1], padding='SAME')\n",
    "    conv_layer1 = tf.nn.bias_add(conv_layer1, bias1)\n",
    "    conv_layer1 = tf.nn.relu(conv_layer1)\n",
    "    \n",
    "    pool1 = tf.nn.max_pool(conv_layer1, [1, 3, 3, 1], [1, 1, 1, 1], padding='VALID')\n",
    "    \n",
    "    print(np.shape(pool1))\n",
    "\n",
    "    filter_size_height = 5\n",
    "    filter_size_width = 5\n",
    "    color_channels = 6\n",
    "    k_output = 16\n",
    "\n",
    "    weight2 = tf.Variable(tf.truncated_normal(\n",
    "    [filter_size_height, filter_size_width, color_channels, k_output], mean=0, stddev=0.1))\n",
    "    bias2 = tf.Variable(tf.zeros(k_output))\n",
    "    \n",
    "    conv_layer2 = tf.nn.conv2d(pool1, weight2, strides=[1, 1, 1, 1], padding='VALID')\n",
    "    conv_layer2 = tf.nn.bias_add(conv_layer2, bias2)\n",
    "    conv_layer2 = tf.nn.relu(conv_layer2)\n",
    "    \n",
    "    print(np.shape(conv_layer2))\n",
    "\n",
    "    pool2 = tf.nn.max_pool(conv_layer2, [1, 2, 2, 1], [1, 2, 2, 1], padding='VALID')\n",
    "    \n",
    "    print(np.shape(pool2))\n",
    "    \n",
    "    filter_size_height = 5\n",
    "    filter_size_width = 5\n",
    "    color_channels = 16\n",
    "    k_output = 24\n",
    "\n",
    "    weight3 = tf.Variable(tf.truncated_normal(\n",
    "    [filter_size_height, filter_size_width, color_channels, k_output], mean=0, stddev=0.1))\n",
    "    bias3 = tf.Variable(tf.zeros(k_output))\n",
    "\n",
    "    conv_layer3 = tf.nn.conv2d(pool2, weight3, strides=[1, 1, 1, 1], padding='VALID')\n",
    "    conv_layer3 = tf.nn.bias_add(conv_layer3, bias3)\n",
    "    conv_layer3 = tf.nn.relu(conv_layer3)\n",
    "    \n",
    "    print(np.shape(conv_layer3))\n",
    "\n",
    "    pool3 = tf.nn.max_pool(conv_layer3, [1, 2, 2, 1], [1, 2, 2, 1], padding='VALID')\n",
    "    print(np.shape(pool3))\n",
    "\n",
    "    flattened = flatten(pool3)\n",
    "    \n",
    "    flattened = tf.nn.dropout(flattened, keep_prob)\n",
    "\n",
    "    weights4 = tf.Variable(tf.truncated_normal(shape=(384,240), mean=mu, stddev=sigma))\n",
    "    bias4 = tf.Variable(tf.zeros(240))\n",
    "    fc1 = tf.add(tf.matmul(flattened, weights4), bias4)\n",
    "\n",
    "    fc1 = tf.nn.relu(fc1)\n",
    "\n",
    "    weights5 = tf.Variable(tf.truncated_normal(shape=(240,120), mean=mu, stddev=sigma))\n",
    "    bias5 = tf.Variable(tf.zeros(120))\n",
    "    fc2 = tf.add(tf.matmul(fc1, weights5), bias5)\n",
    "\n",
    "    fc2 = tf.nn.relu(fc2)\n",
    "\n",
    "    weights6 = tf.Variable(tf.truncated_normal(shape=(120,43), mean=mu, stddev=sigma))\n",
    "    bias6 = tf.Variable(tf.zeros(43))\n",
    "    fc3 = tf.matmul(fc2, weights6)\n",
    "    logits = tf.add(fc3, bias6)\n",
    "    \n",
    "    return logits\n",
    "\n",
    "import tensorflow as tf\n",
    "from sklearn.utils import shuffle\n",
    "import cv2\n",
    "\n",
    "X_train_processed, y_train_processed = process(X_train, y_train)\n",
    "X_valid_processed, y_valid_processed = process(X_valid, y_valid)\n",
    "X_train_processed = X_train_processed[..., np.newaxis]\n",
    "X_valid_processed = X_valid_processed[..., np.newaxis]\n",
    "\n",
    "X_train_processed, y_train_processed = shuffle(X_train_processed, y_train_processed)\n",
    "\n",
    "EPOCHS = 15\n",
    "BATCH_SIZE= 85\n",
    "\n",
    "x = tf.placeholder(tf.float32, (None, 32, 32, 1))\n",
    "y = tf.placeholder(tf.int32, (None))\n",
    "one_hot_y = tf.one_hot(y, 43)\n",
    "\n",
    "rate = 0.0007\n",
    "\n",
    "logits = LeNet(x)\n",
    "cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=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)\n",
    "\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})\n",
    "        total_accuracy += (accuracy * len(batch_x))\n",
    "    return total_accuracy / num_examples"
   ]
  },
  {
   "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": 13,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training...\n",
      "\n",
      "EPOCH 1 ...\n",
      "Validation Accuracy = 0.763\n",
      "\n",
      "EPOCH 2 ...\n",
      "Validation Accuracy = 0.835\n",
      "\n",
      "EPOCH 3 ...\n",
      "Validation Accuracy = 0.888\n",
      "\n",
      "EPOCH 4 ...\n",
      "Validation Accuracy = 0.888\n",
      "\n",
      "EPOCH 5 ...\n",
      "Validation Accuracy = 0.900\n",
      "\n",
      "EPOCH 6 ...\n",
      "Validation Accuracy = 0.914\n",
      "\n",
      "EPOCH 7 ...\n",
      "Validation Accuracy = 0.927\n",
      "\n",
      "EPOCH 8 ...\n",
      "Validation Accuracy = 0.919\n",
      "\n",
      "EPOCH 9 ...\n",
      "Validation Accuracy = 0.930\n",
      "\n",
      "EPOCH 10 ...\n",
      "Validation Accuracy = 0.920\n",
      "\n",
      "EPOCH 11 ...\n",
      "Validation Accuracy = 0.937\n",
      "\n",
      "EPOCH 12 ...\n",
      "Validation Accuracy = 0.929\n",
      "\n",
      "EPOCH 13 ...\n",
      "Validation Accuracy = 0.926\n",
      "\n",
      "EPOCH 14 ...\n",
      "Validation Accuracy = 0.926\n",
      "\n",
      "EPOCH 15 ...\n",
      "Validation Accuracy = 0.941\n",
      "\n",
      "Model saved\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    num_examples = len(X_train_processed)\n",
    "    \n",
    "    print(\"Training...\")\n",
    "    print()\n",
    "    for i in range(EPOCHS):\n",
    "        X_train_processed, y_train_processed = shuffle(X_train_processed, y_train_processed)\n",
    "        for offset in range(0, num_examples, BATCH_SIZE):\n",
    "            end = offset + BATCH_SIZE\n",
    "            batch_x, batch_y = X_train_processed[offset:end], y_train_processed[offset:end]\n",
    "            sess.run(training_operation, feed_dict={x: batch_x, y: batch_y})\n",
    "            \n",
    "        validation_accuracy = evaluate(X_valid_processed, y_valid_processed)\n",
    "        print(\"EPOCH {} ...\".format(i+1))\n",
    "        print(\"Validation Accuracy = {:.3f}\".format(validation_accuracy))\n",
    "        print()\n",
    "        \n",
    "    saver.save(sess, './lenet')\n",
    "    print(\"Model saved\")"
   ]
  },
  {
   "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": 14,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAABpCAYAAAAqXNiiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhAiHCf4/k0gRfR53PV0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd295470b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import cv2\n",
    "import os\n",
    "import numpy as np\n",
    "# Visualizations will be shown in the notebook.\n",
    "%matplotlib inline\n",
    "\n",
    "dirPath = \"test_images/\"\n",
    "files = os.listdir(dirPath)\n",
    "test_images = []\n",
    "\n",
    "plt.figure(1)\n",
    "for file in files:\n",
    "    image = cv2.imread(dirPath + file)\n",
    "    image = cv2.resize(image, (32, 32))\n",
    "    test_images.append(image)\n",
    "    ax = plt.subplot(151 + files.index(file))\n",
    "    ax.set_title(\"Number\"+str(files.index(file)))\n",
    "    plt.imshow(image)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Predict the Sign Type for Each Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy = 0.919\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "    \n",
    "    X_test_processed, y_test_processed = process(X_test, y_test)\n",
    "    X_test_processed = X_test_processed[..., np.newaxis]\n",
    "    X_test_processed, y_test_processed = shuffle(X_test_processed, y_test_processed)\n",
    "    test_accuracy = evaluate(X_test_processed, y_test_processed)\n",
    "    print(\"Test Accuracy = {:.3f}\".format(test_accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy = 0.985\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "\n",
    "    X_test_processed, y_test_processed = process(X_train, y_train)\n",
    "    X_test_processed = X_test_processed[..., np.newaxis]\n",
    "    X_test_processed, y_test_processed = shuffle(X_test_processed, y_test_processed)\n",
    "    test_accuracy = evaluate(X_test_processed, y_test_processed)\n",
    "    print(\"Training Accuracy = {:.3f}\".format(test_accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Analyze Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy = 0.800\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    #sess.run(tf.global_variables_initializer())\n",
    "    #saver = tf.train.import_meta_graph('./lenet.meta')\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "    X_test_processed, y_test_processed = process(test_images, [12, 27, 2, 5, 35])\n",
    "    X_test_processed = X_test_processed[..., np.newaxis]\n",
    "    X_test_processed, y_test_processed = shuffle(X_test_processed, y_test_processed)\n",
    "    test_accuracy = evaluate(X_test_processed, y_test_processed)\n",
    "    print(\"Test Accuracy = {:.3f}\".format(test_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": 20,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Softmax 0\n",
      "[  9.68340397e-01   2.87030190e-02   1.08602922e-03   4.23786987e-04\n",
      "   3.02637636e-04]\n",
      "Top Guesses 0\n",
      "a) 12\n",
      "b) 40\n",
      "c) 38\n",
      "d) 1\n",
      "e) 35\n",
      "Softmax 1\n",
      "[ 0.64811838  0.27712172  0.03389869  0.01506209  0.0149643 ]\n",
      "Top Guesses 1\n",
      "a) 18\n",
      "b) 11\n",
      "c) 26\n",
      "d) 37\n",
      "e) 40\n",
      "Softmax 2\n",
      "[ 0.96986639  0.01685805  0.00793509  0.00349806  0.00167412]\n",
      "Top Guesses 2\n",
      "a) 2\n",
      "b) 3\n",
      "c) 5\n",
      "d) 14\n",
      "e) 1\n",
      "Softmax 3\n",
      "[  9.28102612e-01   7.18947202e-02   1.47685796e-06   6.27514964e-07\n",
      "   3.63612713e-07]\n",
      "Top Guesses 3\n",
      "a) 5\n",
      "b) 3\n",
      "c) 7\n",
      "d) 2\n",
      "e) 4\n",
      "Softmax 4\n",
      "[  9.99709904e-01   2.85126967e-04   3.23948893e-06   4.69287926e-07\n",
      "   3.94240089e-07]\n",
      "Top Guesses 4\n",
      "a) 35\n",
      "b) 36\n",
      "c) 40\n",
      "d) 13\n",
      "e) 33\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAABcCAYAAAB+6068AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhw8fhwn8Ad2HDx8+DhP8f/B30GNgmJ/0AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd291d5da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "softmax_logits = tf.nn.softmax(logits)\n",
    "top_k = tf.nn.top_k(softmax_logits, k=5)\n",
    "\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "    X_test_processed, y_test_processed = process(test_images, [12, 27, 2, 5, 35])\n",
    "    X_test_processed_extend = X_test_processed[..., np.newaxis]\n",
    "    #softmax_logits_out = sess.run(softmax_logits, feed_dict={x: X_test_processed})\n",
    "    top_k_out = sess.run(top_k, feed_dict={x: X_test_processed_extend})\n",
    "\n",
    "    plt.figure(1)\n",
    "        \n",
    "    for i, image in enumerate(X_test_processed):\n",
    "        plt.subplot(151 + i)\n",
    "        plt.imshow(image)\n",
    "        guess1 = top_k_out[1][i][0]\n",
    "        guess2 = top_k_out[1][i][1]\n",
    "        guess3 = top_k_out[1][i][2]\n",
    "        guess4 = top_k_out[1][i][3]\n",
    "        guess5 = top_k_out[1][i][4]\n",
    "        print(\"Softmax\", i)\n",
    "        #print(softmax_logits_out[i])\n",
    "        print(top_k_out[0][i])\n",
    "        print(\"Top Guesses\", i)\n",
    "        print(\"a)\",guess1)\n",
    "        print(\"b)\",guess2)\n",
    "        print(\"c)\",guess3)\n",
    "        print(\"d)\",guess4)\n",
    "        print(\"e)\",guess5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "---\n",
    "\n",
    "## Step 4: Visualize the Neural Network's State with Test Images\n",
    "\n",
    " This Section is not required to complete but acts as an additional excersise for understaning the output of a neural network's weights. While neural networks can be a great learning device they are often referred to as a black box. We can understand what the weights of a neural network look like better by plotting their feature maps. After successfully training your neural network you can see what it's feature maps look like by plotting the output of the network's weight layers in response to a test stimuli image. From these plotted feature maps, it's possible to see what characteristics of an image the network finds interesting. For a sign, maybe the inner network feature maps react with high activation to the sign's boundary outline or to the contrast in the sign's painted symbol.\n",
    "\n",
    " Provided for you below is the function code that allows you to get the visualization output of any tensorflow weight layer you want. The inputs to the function should be a stimuli image, one used during training or a new one you provided, and then the tensorflow variable name that represents the layer's state during the training process, for instance if you wanted to see what the [LeNet lab's](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) feature maps looked like for it's second convolutional layer you could enter conv2 as the tf_activation variable.\n",
    "\n",
    "For an example of what feature map outputs look like, check out NVIDIA's results in their paper [End-to-End Deep Learning for Self-Driving Cars](https://devblogs.nvidia.com/parallelforall/deep-learning-self-driving-cars/) in the section Visualization of internal CNN State. NVIDIA was able to show that their network's inner weights had high activations to road boundary lines by comparing feature maps from an image with a clear path to one without. Try experimenting with a similar test to show that your trained network's weights are looking for interesting features, whether it's looking at differences in feature maps from images with or without a sign, or even what feature maps look like in a trained network vs a completely untrained one on the same sign image.\n",
    "\n",
    "<figure>\n",
    " <img src=\"visualize_cnn.png\" width=\"380\" alt=\"Combined Image\" />\n",
    " <figcaption>\n",
    " <p></p> \n",
    " <p style=\"text-align: center;\"> Your output should look something like this (above)</p> \n",
    " </figcaption>\n",
    "</figure>\n",
    " <p></p> \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2oAAAEUCAYAAABanMWTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n237kWDYT1awAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x17f2ffa9048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def outputFeatureMap(image_input, tf_activation, activation_min=-1, activation_max=-1 ,plt_num=1):\n",
    "    activation = tf_activation.eval(session=sess,feed_dict={x : image_input})\n",
    "    featuremaps = activation.shape[3]\n",
    "    plt.figure(plt_num, figsize=(15,15))\n",
    "    for featuremap in range(featuremaps):\n",
    "        plt.subplot(6,8, featuremap+1) # sets the number of feature maps to show on each row and column\n",
    "        plt.title('FeatureMap ' + str(featuremap)) # displays the feature map number\n",
    "        if activation_min != -1 & activation_max != -1:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", vmin =activation_min, vmax=activation_max, cmap=\"gray\")\n",
    "        elif activation_max != -1:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", vmax=activation_max, cmap=\"gray\")\n",
    "        elif activation_min !=-1:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", vmin=activation_min, cmap=\"gray\")\n",
    "        else:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", cmap=\"gray\")\n",
    "            \n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "\n",
    "    X_test_processed, y_test_processed = process(test_images, [0, 0 , 0, 0, 0])\n",
    "    X_test_processed = X_test_processed[..., np.newaxis]      \n",
    "    outputFeatureMap([X_test_processed[4]], conv_layer2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Question 9\n",
    "\n",
    "Discuss how you used the visual output of your trained network's feature maps to show that it had learned to look for interesting characteristics in traffic sign images\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "**Answer:**"
   ]
  },
  {
   "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
}