Post-Process.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "CSC420 Neural Network",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UMkKZCihZ3I4",
        "colab_type": "text"
      },
      "source": [
        "# Pre-requisites:\n",
        "Please download this folder: https://drive.google.com/open?id=1xyssfNT-zIv7bN5jVhStAGzHLWj3FJSL\n",
        "\n",
        "and put it in your main google drive. This is where some sample images are located and where the master keyword list is (for the OCR step)\n",
        "\n",
        "You can rename it whatever you like, and put it wherever you like, but make sure the folder you downloaded is somewhere in your google drive! Then in the section under 'Python Globals', please set base_dir to point to whatever you renamed it to (if you didn't rename it and you put it in your main directory, then you don't need to do anything about that!)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lnbPJelJ39FO",
        "colab_type": "code",
        "outputId": "0f1e6a83-03b1-4be4-b148-25d4ceb7e645",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "source": [
        "# Mount drive code\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HaYcHxy4USH2",
        "colab_type": "code",
        "outputId": "1acf5aca-8d3d-4b14-bf00-169028c52fb4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 405
        }
      },
      "source": [
        "# Setup the environment\n",
        "!sudo apt install tesseract-ocr\n",
        "\n",
        "!pip install torch\n",
        "!pip3 install torchvision\n",
        "!pip install pytesseract\n",
        "!pip install python-Levenshtein\n",
        "!pip install munkres"
      ],
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "tesseract-ocr is already the newest version (4.00~git2288-10f4998a-2).\n",
            "The following package was automatically installed and is no longer required:\n",
            "  libnvidia-common-430\n",
            "Use 'sudo apt autoremove' to remove it.\n",
            "0 upgraded, 0 newly installed, 0 to remove and 32 not upgraded.\n",
            "Requirement already satisfied: torch in /usr/local/lib/python3.6/dist-packages (1.3.1)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torch) (1.17.4)\n",
            "Requirement already satisfied: torchvision in /usr/local/lib/python3.6/dist-packages (0.4.2)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from torchvision) (1.12.0)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torchvision) (1.17.4)\n",
            "Requirement already satisfied: torch==1.3.1 in /usr/local/lib/python3.6/dist-packages (from torchvision) (1.3.1)\n",
            "Requirement already satisfied: pillow>=4.1.1 in /usr/local/lib/python3.6/dist-packages (from torchvision) (4.3.0)\n",
            "Requirement already satisfied: olefile in /usr/local/lib/python3.6/dist-packages (from pillow>=4.1.1->torchvision) (0.46)\n",
            "Requirement already satisfied: pytesseract in /usr/local/lib/python3.6/dist-packages (0.3.0)\n",
            "Requirement already satisfied: Pillow in /usr/local/lib/python3.6/dist-packages (from pytesseract) (4.3.0)\n",
            "Requirement already satisfied: olefile in /usr/local/lib/python3.6/dist-packages (from Pillow->pytesseract) (0.46)\n",
            "Requirement already satisfied: python-Levenshtein in /usr/local/lib/python3.6/dist-packages (0.12.0)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from python-Levenshtein) (41.6.0)\n",
            "Requirement already satisfied: munkres in /usr/local/lib/python3.6/dist-packages (1.1.2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b9VOer7pO2I4",
        "colab_type": "text"
      },
      "source": [
        "# Python globals"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O4jZ881AA-aU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Python globals\n",
        "base_dir = '/content/drive/My Drive/CSC420 Final Project'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n2YCZZlpamLe",
        "colab_type": "text"
      },
      "source": [
        "# OCR and Post-processing steps\n",
        "\n",
        "## Using Tesseract for Python\n",
        "This part feeds the image into our OCR software (pytesseract in this case), gets the output, and does post-processing on it (removing bad text / lines, and creating key-value pairs using bipartite matching).\n",
        "\n",
        "Please run the two code blocks below, as they handle definitions of the helper functions (and their dependencies)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CKkA_qvxAZK7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Helper functions!\n",
        "from Levenshtein import distance\n",
        "import numpy as np\n",
        "from munkres import Munkres\n",
        "import cv2\n",
        "import pytesseract\n",
        "from PIL import Image\n",
        "import matplotlib.pyplot as plt\n",
        "import re\n",
        "\n",
        "\n",
        "def has_numbers(line):\n",
        "    return bool(re.compile('\\d').search(line))\n",
        "    #return any(c.isdigit() for c in s)\n",
        "\n",
        "def has_letters(line):\n",
        "    return bool(re.compile('[a-zA-Z]').search(line))\n",
        "    #return any(c.isalpha() for c in s)\n",
        "\n",
        "def bipartite_match(pairs, keywords):\n",
        "    # Use the Munkres library to utilize the Hungarian algorithm\n",
        "    # to calculate the min distance matchings (for candidates and true keywords)\n",
        "    munkres = Munkres()\n",
        "    pairs = list(pairs)\n",
        "    keywords = list(keywords)\n",
        "\n",
        "    keywords_length = len(keywords)\n",
        "    pairs_length = len(pairs)\n",
        "\n",
        "    # Create distance map to feed into munkres.compute function\n",
        "    distances = np.zeros((keywords_length, pairs_length)).astype(np.int32).tolist()\n",
        "    for i in range(keywords_length):\n",
        "        for j in range(pairs_length):\n",
        "            candidate = pairs[j][0]\n",
        "            distances[i][j] = distance(candidate, keywords[i])\n",
        "    \n",
        "    # Compute possible matchings, and then filter out the bad ones\n",
        "    matchings = munkres.compute(distances)\n",
        "    #print matchings\n",
        "\n",
        "    good_pairs = []\n",
        "    for i, j in matchings:\n",
        "        keyword = keywords[i]\n",
        "        distance_threshold = len(keyword) / 2\n",
        "        if distances[i][j] < distance_threshold:\n",
        "            # If matching is considered 'good' enough,\n",
        "            # then we use it. Otherwise it is discarded\n",
        "            good_pairs.append((keyword, pairs[j][1], pairs[j][2]))\n",
        "\n",
        "    return good_pairs\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NeQl4BChKwGb",
        "colab_type": "text"
      },
      "source": [
        "# Second helper function block"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WvTApXJLeR-G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# More helper functions. Please run this as well\n",
        "def create_pairs(lines):\n",
        "    pairs = []\n",
        "\n",
        "    # Modify in place the lines\n",
        "    # Get rid of empty lines\n",
        "    lines = filter(None, lines)\n",
        "    lines = filter(lambda line: line != ' ', lines)\n",
        "\n",
        "    for i, line in enumerate(lines):\n",
        "        # Skip the line if it doesn't seem like a good line\n",
        "        if (len(line) <= 5 or (not has_numbers(line)) or (not has_letters(line))):\n",
        "            continue\n",
        "\n",
        "        # This step picks some possible candidate keywords.\n",
        "        # They need to be at least as large as some threshold\n",
        "        count_threshold = 3\n",
        "        letter_count = 0\n",
        "\n",
        "        for j in range(len(line)):\n",
        "            if (has_letters(line[j])):\n",
        "                letter_count += 1\n",
        "            # Create the candidate keyword\n",
        "            if (letter_count > count_threshold and has_numbers(line[j])): \n",
        "                before_space_index = j - 1\n",
        "                # Likely will have a space, so remove it\n",
        "                line = line[0 : before_space_index] \\\n",
        "                  + ' ' \\\n",
        "                  + line[j : len(line) - 1] \n",
        "                \n",
        "                break\n",
        "\n",
        "        # Look for a rough value string to go with the candidate keyword\n",
        "        line_words = line.strip().split(' ')\n",
        "        key, words = line_words[0], line_words[1:]\n",
        "\n",
        "        for k, word in enumerate(words):\n",
        "          # First occurence with a number is likely to be\n",
        "          # a value (such as sodium amount, calorie content, etc)\n",
        "            if has_numbers(word):\n",
        "                words = words[k : ]\n",
        "                break\n",
        "            # If not a number, might be part of the nutrient keyword name\n",
        "            # so add it to the candidate key\n",
        "            else:\n",
        "                key += ' ' + word\n",
        "        \n",
        "        # Rough value at best\n",
        "        # TODO: clean-up further\n",
        "        value = ('').join(words)\n",
        "        pairs.append((key, value, i))\n",
        "\n",
        "    # Cleanup step 1:\n",
        "    # Remove obviously bad pairs created initially\n",
        "    good_pairs = []\n",
        "\n",
        "    # Cleanup the pairs we created by doing some Levenshtein\n",
        "    # distance measurement betwen keywords and candidates to \n",
        "    # get rid of obvious non-keywords\n",
        "    for pair in pairs:\n",
        "        candidate_len = len(pair[0])\n",
        "        is_good = False\n",
        "\n",
        "        # Check part of the candidate using a window; if the distance is below\n",
        "        # some threshold for some keyword, then it's probably a good candidate\n",
        "        for key in keywords:\n",
        "            key_len = len(key)\n",
        "            window = candidate_len - len(key) + 1\n",
        "            #print('Window:', window)\n",
        "            #print(pair[0], key)\n",
        "            if window <= 0:\n",
        "                window = 1\n",
        "            \n",
        "            for i in range(window):\n",
        "                threshold = key_len / 2 #key_len / 3\n",
        "\n",
        "                dist = distance(key, pair[0][i : key_len + i - 1])\n",
        "\n",
        "                if dist <= threshold:\n",
        "                    is_good = True\n",
        "                    break\n",
        "\n",
        "        if is_good:\n",
        "            good_pairs.append(pair)\n",
        "\n",
        "    return good_pairs\n",
        "\n",
        "# Handle matching keywords with candidate pairs\n",
        "def match_pairs(pairs):\n",
        "    new_keys = []\n",
        "\n",
        "    for i, k in enumerate(pairs, 1):\n",
        "        new_keys.append((k[0], k[1], i))\n",
        "\n",
        "    key_pairs = bipartite_match(new_keys, keywords)\n",
        "\n",
        "    # Check correct ordering. Remove candidates where the line # is too large\n",
        "    # TODO: confirm where the bad index is coming from for File_002.png\n",
        "    bad_indices = []\n",
        "\n",
        "    for candidate in key_pairs:\n",
        "        line_num = candidate[2]\n",
        "        if line_num > len(key_pairs):\n",
        "            #print(candidate)\n",
        "            bad_indices.append(candidate[2])\n",
        "\n",
        "    for i in bad_indices:\n",
        "        for tup in new_keys:\n",
        "            if i == tup[2]:\n",
        "                new_keys.remove(tup) # Remove extraneous lines\n",
        "\n",
        "    #print(bad_indices)\n",
        "\n",
        "    # Too many lines, means we need to re-run the\n",
        "    # matching again on the new, trimmed list of candidates\n",
        "    if len(bad_indices) > 0:\n",
        "        key_pairs = bipartite_match(new_keys, keywords)\n",
        "\n",
        "    return key_pairs\n",
        "\n",
        "\n",
        "def clean_values(tuples):\n",
        "    # Gets rid of letters and non-numbers (excluding the period)\n",
        "    # so we can convert to float later, for displaying\n",
        "    # into a pie graph\n",
        "    clean_tuples = []\n",
        "    for tup in tuples:\n",
        "        tup_copy = list(tup)\n",
        "        value = tup_copy[1]\n",
        "        value = re.sub('[^0-9.]', '', value)\n",
        "\n",
        "        tup_copy[1] = value\n",
        "        clean_tuples.append(tuple(tup_copy))\n",
        "\n",
        "    return clean_tuples\n",
        "\n",
        "# Creates the mapping we'll use to chart the pie graph\n",
        "def post_process(ocr_text):\n",
        "    # Create candidate key-value pairs\n",
        "    lines = ocr_text.split('\\n')\n",
        "    candidate_pairs = create_pairs(lines)\n",
        "    #print('Candidate pairs:\\n', candidate_pairs)\n",
        "\n",
        "    # Use bipartite matching to match each true keyword\n",
        "    # to a candidate pair\n",
        "    true_pairs = match_pairs(candidate_pairs)\n",
        "    #print('True pairs:\\n', true_pairs)\n",
        "\n",
        "    # Perform final cleanup of the values in the key-value pairs\n",
        "    cleaned = clean_values(true_pairs)\n",
        "    #print('Cleaned up pairs:\\n', cleaned)\n",
        "\n",
        "    key_mapping = {}\n",
        "    for p in cleaned:\n",
        "      key_mapping[p[0]] = p[1]\n",
        "    print('Final mapping:\\n', key_mapping)\n",
        "\n",
        "    return key_mapping"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ITCMKqPSIX8u",
        "colab_type": "text"
      },
      "source": [
        "# Displaying the pie chart\n",
        "If you run this cell, you should get a printout of the final detected nutrients and their values after postprocessing.\n",
        "\n",
        "Alongside this, you will see a pie chart (which admittedly isn't displayed very well; something to put on the bucket list to fix). Below that you will see a picture of the food label it is based off of.\n",
        "\n",
        "If you want to try different photos, please put them inside the <i> ${base_dir}/input_dataset </i> where base_dir is as as defined in the third code block. Then modify the image_path variable below to point to your image. \n",
        "\n",
        "Due to my post-processing step being not perfect, sometimes the code may bug out. I apologize for this. It currently works for certain with the stock image provided below."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J4lL5B9BFcTx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 620
        },
        "outputId": "c61c1638-ddb1-4713-ad76-6e0d7a0b7130"
      },
      "source": [
        "# OCR and Main post processing step\n",
        "# The above two code cells need to be run first\n",
        "# The two variables below can be modified to put in whatever you like to try.\n",
        "nutrients_path = base_dir + '/na_nutrients.txt'\n",
        "image_path = base_dir + '/input_dataset/File_002.png'\n",
        "\n",
        "img = cv2.imread(image_path)\n",
        "\n",
        "# Create a set of known nutrient keywords we're searching for\n",
        "keywords = {}\n",
        "nutrients_file = open(nutrients_path, 'r+')\n",
        "for line in nutrients_file:\n",
        "  keywords[line.strip()] = \"\"\n",
        "\n",
        "print('Keywords loaded:\\n', keywords)\n",
        "\n",
        "# Post process\n",
        "output_string = pytesseract.image_to_string(Image.open(image_path))\n",
        "results = post_process(output_string)\n",
        "\n",
        "# Display calories separately, not part of the pie chart\n",
        "# Some hacky code...\n",
        "calories = ''\n",
        "if 'calories' in results or 'calories/calories' in results:\n",
        "  results.pop('calories', None)\n",
        "  results.pop('calories/calories', None)\n",
        "if 'calories' in results:\n",
        "  calories = results['calories']\n",
        "\n",
        "if 'calories/calories' in results:\n",
        "  calories = results['calories/calories']\n",
        "\n",
        "keys = results.keys()\n",
        "values = [float(results[k]) for k in keys]\n",
        "\n",
        "# Credit for absolute_value function goes to ImportanceOfBeingErnest\n",
        "# from StackOverFlow:\n",
        "# https://stackoverflow.com/questions/41088236/how-to-have-actual-values-in-matplotlib-pie-chart-displayed-python\n",
        "def absolute_value(val):\n",
        "    a  = np.round(val/100 * sum(values), 0)\n",
        "    return a\n",
        "\n",
        "fig1 = plt.pie(values, labels=keys, shadow=True, autopct=absolute_value, startangle=90)\n",
        "#plt.legend()\n",
        "plt.title('Calories: ' + calories)\n",
        "\n",
        "plt.show()\n",
        "plt.imshow(img)"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Keywords loaded:\n",
            " {'calories': '', 'calories/calories': '', 'fat/lipides': '', 'fat': '', 'saturated/satures': '', 'saturated': '', 'trans/trans': '', 'carbohydrate': '', 'cholesterol': '', 'cholesterol/cholesterol': '', 'sodium/sodium': '', 'sodium': '', 'carbohydrate/glucides': '', 'fibre/fibres': '', 'fibre': '', 'fiber': '', 'dietary fiber': '', 'sugars/sucres': '', 'sugars': '', 'protein': '', 'total fat': '', 'total carbohydrate': ''}\n",
            "Final mapping:\n",
            " {'calories': '80', 'fat/lipides': '9', 'saturated/satures': '1.5', 'trans/trans': '0', 'carbohydrate/glucides': '0'}\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQAAAAD3CAYAAADykopzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO2dd3hc1bW33zVN0kga2ZZ7FdgDxrjR\nIlooxmAwnZhrepyEBBxIgTR/CdwI0iBcCKFfkptAQgkECAE79I7BAhtXjI3AvSHb6hqVKev7Yx8Z\n2VjNSDpT9vs882jm1N8ZzfmdXdZeW1QVi8WSmXjcFmCxWNzDGoDFksFYA7BYMhhrABZLBmMNwGLJ\nYKwBWCwZjDWAFENEVETG7OO+F4vIi92tyZK6WANwARG5SEQWikidiGwVkedE5NiePq+qPqyqp/T0\neSypgzWAXkZErgVuB34LDAJGAvcAZ/fweX09eXxLamINoBcRkQLgRuAqVX1KVetVNaqqz6rqT5xt\nviIi74pIlVM6uEtEAm0dT0T+JiLbRWS9iFwnIh5n3SwRmS8ifxCRnUCJs+ztVvuPFZGXRKRCRFaL\nyH+1WjddRFaKSK2IbBaRH/fol2NxBWsAvctRQDbwr3a2iQPXAP2d7U8CvtvGtncCBcD+wPHAZcA3\nWq0vBtZgShq/ab2jiOQCLwGPAAOBC4B7RGScs8n/AVeoaj4wHni11b5VvVFlsfQ81gB6l0Jgh6rG\n2tpAVRep6gJVjanqOuB/MTf3boiIF3PT/j9VrXW2vRW4tNVmW1T1TudYDXsc4gxgnar+1Vm/GHgS\nON9ZHwXGiUhIVStV9YNWGvuo6ttYUh5rAL3LTqB/e/VxETlAROaKyDYRqcG0FfTfy6b9AT+wvtWy\n9cCwVp83tqNlFFDsPM2rRKQKuBgY7Kz/GjAdWC8ib4jIUR1dnCX1sAbQu7wLNAHntLPNvcAqIKyq\nIeDngOxlux2Yp/SoVstGAptbfW5vqOdG4A3nad7yylPV2QCq+r6qno2pHjwNPN7+pVlSEWsAvYiq\nVgP/DdwtIueISFBE/CJymoj83tksH6gB6kRkLDC7jWPFMTflb0QkX0RGAdcCD3VSzlzgABG51NHg\nF5EjROQgEQk4MQMFqhp19CT2/cotyYo1gF5GVW/F3KjXAdsxT+KrMU9ZgB8DFwG1wJ+Ax9o53PeA\nekxD39uYBr2/dFJHLXAKph1hC7ANuBnIcja5FFjnVEOuxFQPAHDiF77amfNYkhuxCUEslszFlgAs\nlgzGGoDFksFYA7BYMhhrABZLBmMNwGLJYKwBWCwZjDUAiyWDsQZgsWQw1gAslgzGGoDFksFYA7BY\nMhhrABZLBmMNwGLJYHrUAJwklHd1YfsiEVnxJc53gojM7eI+RSJyuYi0lXdvz+2fE5Hh7axv95pF\nZJGIZInIz/ey7kYRmbqX5V2+LoulM/SYASRTGuoOtBQBF7KXxJt77iciOUChqm7aRx37AZtVtQmT\n6Wc3VPW/gVdaMvtaLD1Np35oInKZiCwTkaUi8ncROVNESkVksYi8LCKDnO1KnPXzgb87u48QkddF\npExEftnqmNeKyArn9cNWp/OKyJ9E5EMReVFEckRktIh80GrfcMtnETlVRFY5n89rtc1uWpwn/Vsi\n8oHzOtrZ9Cbgq5gEmJtF5H0R2SQilUC9iFwhIk+LyCJMqq5K5/jTRSQhJnX3dhGpbPkegNHOdVWI\nyMpW1/YR8LyI3AQERaTB2e9eJy33J8AG5zub66xvwCTqaLmu9c73Xu+s/66z/OutljWIyOTO/G8t\nGY6qtvsCDgY+Bvo7n/sBffk8mcjlwK3O+xJgEZDjfJ4FbMVkw80BVgCHA4cBy4FcIA/4EDgE8zSO\nAZOd/R8HLnHev9Zq+W8x2XCyMRl1wpi8eY8Dc9vQEgSynfdhYKHz/gTgFWCF8/l/gGZgP0x2nIWt\nzns3sBYY6pxXgTOBR4FPMFl+ZmGSfw4DHgAua/VdxjApvE/DpP8OOssnYVJuzQNmONe12dHpBcox\n+fsAqoAHnfc/Aqqd73EtcIOzvC+Q19H/1r7sqzPF9CnAP1V1B4CqVojIBOAxERkCBJwfXwvP6O4p\nqF9S1Z0AIvIUcKxz4/xLVetbLf8q8AywVlWXOPsuwpgCwJ+Bb4iZWWcm8BVgrLN9mXOch4DvtKHF\nD9zlPBnjwAFtXO8RmBu1JXd/AfAjEZmIuSGjmFz9azAJM+dizOcAR+smjOk9gDGd1x1tAYxprhGR\nq4Coqkacc1RjMvpudz6PBSKYdGA+jBHlO+tyMNl8lzjHz8IkA10A/MS5vlvUpu22dIJ9rWveCdyl\nqhOAKzBPrBbq99h2z5xjHeUga2r1Pg67TOpJzJPzDGBRi6l0QGst1wCfYZ62h2OMqy2Wq+pkVZ2M\nmWhjP0w7wYvAYj7PmxdV1Zbr0VZaX8eUBnKA34tIIXBcF7QOBYYDJ6nqREwppPX/6uuOtouABar6\nkapeiEnlDfCaiHy9g/NZLJ0ygFeB850fMSLSD/NUbEk/3dEP7WQR6ec0oJ0DzAfeAlqy4uYC5zrL\n2kRVG4EXMGmz/+osXgUUicho5/OF7RyiANiqqglMwkuvs7wWY2AtT9j3gVEi4nc+H4TJinsi5uY/\nElNP35/P03XP3ONcIVUtBZ7CmNgI4Co+/75fAnwiEmqlrTXbHX2FTrvCMa3WNQDfFpGWc+eB6SkA\nnleTynsNHRuOxdJxFUBVPxSR3wBviEgccxOUAP90GspexTwh2+I9zNN7OPCQqi4EEJEHnHUAf1bV\nxSJS1IGchzFm8aKjrVFEvgPME5EIxkTy29j3HuBJEbkMeJ7Pn7jLMKWOkIhsxRTha4EPnJusJf/+\nTZi8/gswbQTfBf7tNA6+jymyt3CMiCzn8+/3QUx1od7R/byYOfq2i0g9pg1hF6r6voi8AawEGh1N\nLVRjqjPLMFWAPs7y3wKHiohiGip/0daXaLG0kFJZgcVMUFmgqtf38nmzgPmqenirZXmqWueYxN1A\nmar+oY39hwN/UtXTekexxdI5UsYARORfwGhgSkuDpMt6rsFUfwKYUtG3WzXqWSwpQcoYgMUCJtIS\neFFVt3TT8YqAo1X1kS7u9wCmy/kJ5/MFwGhV/U27O+5+jD7ARap6T1fO3Z3YiDNLqjEL00vSaaTj\nSNCLvoSeFk7DtC11hT60PfV7m4iZGbpbsAZgcR0RyRWReWIiTVeIyEwR+W8nKnOFiNwvhhmYLtyH\nRWSJEyW6TkT6O8c5XERed953KRLUOd41IuIVkVuccy8TkSuc44mI3OVEbL6MadRt0S/AZEzD8fHO\nsZY4EZv5IpInIq84510uIme3OvdoZ9tbZI8xH875Zjnv14nIzWIiXs8XEx37vJixJW+JmUcSETnf\n+c6WisibHX33SROvb+mYojnzhmGCkQY6rwF7/C3ExCh4Ad/SrG+vK5D64ZhejJjztxoTxbjJ+fv5\nq6R6O+5wKrBFVU8HEJECTADZjc7nvwNnqOoTInI18ONWvUntHXcccKyqNohIEDjZ6TkKY3peDgfm\nOMc7wzned4BqVT2ipfFXRF7ERKoe6BxzEKaHpmUexkOApaqqTkP1Vao6X0TyML04AOeqao1jVgtE\n5Bnn3OOdmI6Wrtz22KmqhzrbvgJcqaplIlKM6eWagpl8dpqqbnaqGO1iDSAJKZozbwAw3nkdjPmR\nHNyZf2hrmqJRD4EuFJdLClq6HDdgGjbfAeZTUv1ZV867DywHbhWRmzH16rdE5Gsi8lNMV2c/TLj4\ns1087r5Egp4CTHRKG2BiNMKYuIpH1czKvEVEXm21z6nAc877+cBtIvIw8JSqbhITU/JbETkOE/I9\nDGMiXeUxMD1QwNGYrviWdS3BafOBB0TkcUwcSrtYA0gCiubMG4xx7ymqOkXMqMHd6OBJt1eaYxps\nN97xi2QDY5zXFMxYAygpWEOLGZi/Kyip7rbpwlX1YxE5FJgO/Np5ul0FHK6qG0WkhN2jTVsT4/Oq\n7J7btBUJ6uHzJ/OeCPA9VX1ht4Ui09u5hFNwojBV9SYRmedcy3wRmYYJHhsAHKaqURFZ18b1tL6W\n9q7HA1S1lBxao6pXOiWC04FFInJYe1Gz1gBcoGjOvBAwFXPDn9RSf4N9u9HbwiMS66ZD7e+8LnE+\nV1NSUIoZRPUYJdXrv8zBRWQoUKGqD4lIFWaAGcAO52k3A3jCWVbL7sFe6zCDy57j81DovVEAbFLV\nhJgw6daRoK2P9wIwW0RedW7WAzBRr28CV4hIS1DXicAjTnXF12q8y2hVXQ4sF5EjMOM6CoBy53gn\nAqPaOPd6zKjULEwY+UmYad93w6lKrBWR81X1n04bxERVXeqcvxQoFZHTMFGo1gDcxrnpz9JEYiYi\np4gZHNStN3wvUoB56p0C3ERJwTuYOvXj+9iOMAG4RUQSmHaK2Ziw8RXANkykZQsPAPeJGSZ9FHAD\n8H8i8iucgVdt0F4kaFxEljrH/iOmZ6AlEnS7o+VfmFLRSkwV6V1n/5OBl1ud54fOTZ7AVFuew9zk\nz4qJDl2ICWFHVXeKyHwxSXCeU9WfOEX3FZgBdovbuZ6LgXtF5DpM9eYfwFLne2wZHfuKs6xNbBxA\nD1I0Z54PmKaJ+DcQzxmOs/ca7+isHUNzmvv34iljmB/do8BTlFTX9uK5XUFE/owJZV/gtpZ9wRpA\nD1A0Z95gTcR/AHxbPN5Ct3S4YACtacTkN/gbMLc72wws3Yc1gG6kaM68CYnmxuvFHzhXxON69cpl\nA2jNx8BtwIOUVLfV+GZxAWsA3cConz5zmsaj13n82Ud3vHXvkUQG0EI5cAdwFyXV1W6LsVgD+FKM\n/NFT54jILeILjHFby95IQgNooQr4A3A7JdU1bovJZKwB7AMjvv/IkeLx3ufJzpvktpb2SGIDaKES\nuBW4IxMaDJMRawBdYOQ1jxdpPH6fJyd/Wip036WAAbSwBfg+JdVPui0k07AG0AlGXvtEvkYb/+DJ\nCX1dPF7XG/c6SwoZQAvPAldRUr3RbSGZgh0N2AFDL7/nQpD13ty+30qlmz9FORNYSUnBNZQUdNuQ\nV0vb2BJAGwya+etCX98h//D3GfyFqbpShRQsAbRmEfAdSqo/6HBLyz5jSwB7YcjX/3Bp1tADP03l\nmz8NOAx4j5KC2ygpyHVbTLpiSwCtGHj+Df38hcMe9RUMPiUVGvk6IsVLAK1ZA5xLSfUyt4WkG7YE\n4DDwa9dPyxp64Mf+PkPS4uZPM/YH3qWk4AK3haQbGd+oFQwXe/MPPfN3Ofsfdo14/Rn/fSQxQeBR\nSgoOB35GSXXcbUHpQEaXAPImntyn4JiLXszZ75Cf2Js/ZfgR8AIlBa4NskonMtYA+hx/2fiCo2d+\nkDV4zBS3tVi6zEnAQkoK7BToX5KMNIB+U79zRv6k097y9xnS3pRmluSmCHiHkoLuSOmdsWSUAQTD\nxZ6+J3zjqrwJU//hDYa6lGDTkpTkAA9TUtCrU8WlExljAMFwsS9rxIRf5B96+i2erFzbr5xe3EhJ\nwX+7LSIVyQgDCIaLA9lFh/w6/9Dpv/AEcnLc1mPpEW6wJtB10t4AguHi7JzRR9yaP/nUaz2+rF7N\nyWfpdawJdJG0NoBguDg3Z/QRd+RNOHm2eP1+t/VYegVrAl0gbQ0gGC7Ozxo69pa8CVNniddnR5Zl\nFtYEOklaGkAwXJzt7zf8l/mHnTXLPvkzFmsCnSDtDCAYLvZ58wt/GDry/G97Atm2wS+zuYGSgqvd\nFpHMpJUBBMPFHsnKnVVw1MxrvTn5Ibf1WJKCP1BSYId1t0HaxL8Hw8WC13d2n6Nm/tKX33+A23p6\nk6p3/kH1/H8ASvaoSQz6rxt3W1/TmGD8vfVsrVOyvPD8JUGOHZk2//qO8AGPU1JQTEl1mdtiko10\nKgF8Nf+Q02/wFw4f7raQ3iQRa6Z6/qMMnPkrRvzgURo3LKe+rHS3bb4zt5G8gBC9PsQ3D/Fz0ZMN\nbRwtbekLPPvc9VNtqXAP0sIAguHicdmjJl2XPXLiBLe19Db1K17BEwiSM3ICnkCQ7BHjqftg7m7b\nvLwmxg+PNG2hvz85i001SiKROTN1qdJ8b+zM8tnRa/5eNGeeTfbQipQ3gGC4uJ83NOAneZNOPTYT\nE3lEKzbjyfn8weYrGES8vmK3beqa4ZDBpic02+fB64GyiszIBFXdLE0XVl3ZeHPswq8CZwFz3NaU\nTKS0AQTDxT483isKimec6vFn2RZ/y26U1WbXntL4O++CnONaF/1/VTRnnh0C7pDSBgCclX/omRf6\nQgMGuy3ELfz9hpFo+Hx2rVj1Z3hz++22TV4AFm8zCXQaYwniCQj367i0tLE6wYkP1jPu7joOvqeO\nPy5o+sI2qsr3n2tkzB21TLy3jg+2Jkeinrk7htVM996T91ngC62dXuCBojnzbHsAKWwAwXDx+OxR\nk2Znj5yQcfX+1uSOP4lEU4SGDctJNEdo3LiCvEOm77bNlP183L4gCsBPX2piWEjweDr+1/s8cOsp\n2ay8Ko8F38rl7vejrNy++w3+3CcxyirilH0vj/vPzGb2PHcbGOMJTdxYfnzd1Xm3hKKe7LZcbgRm\nbsKMJyUNIBguLvRk5/0gb+K0IzOx3t8ajy9AwdEzKX/sOjb+8UKyhx9M7gFHse3hn/Hr1+qCAPef\nmU1Nk+L/VQ1/WRzl7+d2rrY0JN/DoUNM20F+lnDQAA+ba3ZvO/j3qhiXTQwgIhw53EdVI2ytdaeB\nsbpZGi+svDL6l9AVeZ3Y/JtFc+ad1uOikpyUSwseDBd7gB8XHHPRrKzBYw5yW08y051pwddVJTju\nr/Ws+G4eoazPTfeMRyLMOTawK67gpL/Vc/PUbA4f2rvDL1bV5tRepiU55YERXQlw2AwcvO6m0zN2\nqvJULAEckzX84CmBQaPtzd9L1DUrX3s8wu2nZu928ycLT+8YUXOG9568Lt78AMPI8KpAShlAMFxc\niNd3Wf6kaUdketG/t4jGzc1/8QQ/5x30xXFVw/KFjdWflyI31SjD8nvnfxNLEP9l+Yn1P8y7ORTz\n7LMzfaNozrzpHW+WnqSMAQTDxQJcmD9x2iRPdl6/DnewfGlUlW8908hB/b1ce9Tec6mcdaCPvy1r\nRlVZsClGQZZpO+hpqpqkYWbld2MPhr7dHend/lg0Z15GjhpNpYDw8d7QwBOyiybbVNC9xPyNcf6+\nLMqEgR4m31cHwG9PymKD88S/8vAA08M+/lMWY8yddQT9wl/P7vlwjI9qcmov5Vc5O3KHdtfvdwww\nG7ijm46XMqREI2AwXJwN/KbP8bOmB/qPHOu2nlQhjeYG3MWTO4pqfhosyY97At1dz9gJjM60BsFU\nqQJM9fcftZ+/cMSBbguxuEMsQfwX5VPrf5T321AP3PwAhcDPe+C4SU3SG0AwXFwAnJU38eRxYlv+\nMpLKJk/D+ZVXxx4OfbOn07l/v2jOvJE9fI6kIukNADg5MGj0QH/fofbpn4GsqAnWTm2+JbA49+je\nyOicDfy2F86TNCS1AQTDxf2AU/MmnDzebS2W3uex7fvXnOW7L3+nf0hvRhVdVDRn3sG9eD5XSWoD\nAKZlDTtogK9g4Gi3hVh6j2ic2M/KT63/Wf6vQwlPr3dUCXBtb5/ULZLWAILh4gHA1NxxJ9infwax\ns9ET+Vr1D+OPhS5zc/q2i4vmzBvk4vl7jaQ1AGC6v3BEvi80YH+3hVh6h2XVuTVTo7dmLQt+xe0Z\nnLKAq1zW0CskpQE4df/jgmO/OsptLZbe4dHtY2rO8d8bqvQPSpZJXGYXzZmX9klmktIAgGLxZ3kD\nA0ZNcluIpWeJxon9uHx65P/l3+hGfb89+gOXuS2ip0k6AwiGi/3AabljjxskXr/bRUFLD7Kj0RM5\nt/raxBOhS4Jua2mDa9I9iWjSGQAwDsjPGnGwjflPY5ZU59VMbb4ta0Xw8IDbWtrhQOB4t0X0JMlo\nANOyhh2U580JZWyev3Tnoe0H1JwbuCdUFRiYLPX99rjUbQE9SVIZQDBcPBQ4KGf0EbblPw1pjhO7\ntvyMyHX5JSGVpKrvt8eMdG4MTCoDAI5FJO7vO9SO+Esztjd6ImdX/zjxVOiiZK3vt0UIM59AWpI0\nBhAMF3uB47JHTsoWXyDVfiSWdlhUlV8zNfrHrI+ChyZzfb89/sttAT1F0hgAUAQEs4aPG+O2EEv3\noKo8uH1s7Yysu0PV/sJUqO+3xWlFc+a5GZnYYyRTRWwyEPf3G2aL/2lAc5zoT3eeE306NDPfbS3d\nQA5wBvCY20K6m6QoATipvo/NGjo24AnkFLitx/LlKG/w1p9V81N9OjQznapyaZk4NFlKAMOBguyR\nE/dzW4jly/F+VUHNt7y/zqvJKUyKh0s3cqLbAnqCZPknTQTU129o2G0hln1DVfm/8nG152fdHarx\np93NDzCiaM68tPt9ul4CcNJ9H+PJyY94svMzYghmutEUI/qjiq9F54ZmpEN9vz2mAGVui+hOksGp\n+wCDskeML7Q5/1KPbQ3e+jNr5zA3NCOd6vttkXbTiieDAYwC1F84MqOSMaYDpZV9aqbG78j5OGdi\npkyqcUK6DQ5KBgM4EIh7QwNGuC3E0jlUlfvLx9fOzL4rVOfrmwy/od5iIJBWGapcbwMAJuPx1niD\nBUPcFmLpmMYYzddUnh9/LnReutf32+IIYLnbIroLVw0gGC4OAoMCg8c0i8ebKcXIlGVrg7f+0saf\nBT7JH5+2g2M6QVplDHa7BDAM0EDhSDv0N8l5t7JvzeX+X+fV52RUkX9vpJUBuP3PHAF4vHn9Cl3W\nYWmDhKreWz6x7sLsO0P1mVXfb4u0agNw+x+6H9DgyQlZA0hCGmI0z95xQcPNoTl5iNs/laRhWNGc\neWkTru52FWA4EPFk5/ZzWYdlD7ZEfPUXN83xr80flwn9+11lPDDfbRHdgWu27kQADgEaPYFgX7d0\nWL7I25X9aqYm7sxZmzMuVcfv9zRp0w7gZrkuFwh4QwOyxeuzP7QkIKGqd5UfUntJzl2hiK/Alvnb\nJm2mqnOzCtAPSPj7DrXF/ySgIUbz96ouir8cOjNT+/e7wkC3BXQXbhuAePP72+K/y2yq99ddEv15\n1rq8AzO5f78rpI0BuFnM6w/gCeTYH52LvFHRv2YqdwXXZR9oA7E6zwC3BXQXbpYABgNR8WXZ2X9c\nIKGqd20/vO620I9CbmtJQdKmBOCmAeQDMfEFrAH0MpEoTVdVXaqvhabb+v6+kTYlALd7AawB9DIb\n6v11p0Zu8LyWPz3bbS0pTLBozrw8t0V0B24aQBCIic9OANpbvFYxoOZk7sndkBW29f0vT1r0XrlZ\nBcgF4uL12ydRD5NQ1du3f6X+jtA1tr7ffaSFibppADlArQ0C6lnqozR9t3qWvhGalhZF1iTC7TD6\nbsHNiwgCVUBapVhKJpY2DJDfeK/0bMobnRZPqyQjlWc62oUrBuDMA+gDEppIxNzQkAnMDt5iR1n2\nHGnx4HKrEVCdF2gi7pIGi+XLkBYPLlcMIFJWmgASgNgSgCVFSYsHl5vdgFHAg8bT4ou0ZBxRtwV0\nB+4bgC0BWFKTKrcFdAduGkAM8GjClgAsKUfzuptOr3ZbRHfgdglANN7c5KIGi2Vf2O62gO7CbQPw\nJJoidS5qsFj2hXK3BXQXbhpALeBPNNTWuqjBYtkXrAF0AzuBQKKhxhqAJdWwVYBuYAcQiNVsT4vG\nFEtGYUsA3cAOwBer3FKtquqiDoulq6xzW0B34aYBVAMJjUfjGmuy1QBLKvGh2wK6CzcNoApnPECi\nKVLpog6LpauscFtAd+GmAVS2nD9eV7HVRR0WS6dR1e3rbjrdtgF0A3VAE+CLVW2zBmBJCUQkbYr/\n4KIBRMpKFVgD5Ee3r9vilg6LpYtYA+hGVgG5zdvX7tR4rNllLRZLZ0ib+j+4bwAbAEFVEw21n7ms\nxWLpDB+4LaA7cdsAdhX9Y/UVthpgSWpUtQZY5LaO7sRtA9iJGRTki1VutQZgSXbeXHfT6Wk1fN1V\nA3BSg30K5DVuXL7GBgRakhkRedVtDd2N2yUAgI+A/HjN9rpEY+02t8VYLO1gDaAHWNXyJlqxucxN\nIRZLW6jqTmCZ2zq6m2QwgHU47QDNW1ZbA7AkK6+tu+n0tKujum4AkbLSGLAE6Ne4cfkmjUUb3NZk\nseyJiLzgtoaewHUDcFgEZKOqsZryT90WY7G0RlVjwNNu6+gJksUAdhX9m8vXfuymEIvlC2jirXU3\nnb7DbRk9QVIYQKSstArYBORHyhassmHBlmRCPN5H3dbQUySFATi8C/TR5kg0Wrk5rQZcWFIXVW0G\nnnBbR0+RTAawCDPjqjSuW7LYbTEWCwCJ+HPrbjo9bRPWJI0BRMpKy4GPgb6N65duTDTV73Rbk8Ui\nXt9f3dbQkySNATi8AuQDNG/7dInLWiwZjibi24H/uK2jJ0k2A1gONAP++o/fWWKzBVtcRfWOdTed\nnhazALdFUhlApKy0EXgbGBCvKa+LVZfbLkGLK2gi3iRe391u6+hpksoAHOYDfoCGT0rfcVmLJUPR\nePThdG78ayEZDWAtJlFIQeP6JRvi9ZWb3BZkySxUVT3+7N+5raM3SDoDcJKFPg30BWhY+8Hb7iqy\nZBoaa3p53U2nf+K2jt4g6QzAYQlm3oBgZPX81XGbL9DSi4g3cKPbGnqLpDSASFlpFHgGGADQsGbh\nG+4qsmQKieaGBet/f2bGlDqT0gAc3gXqgZzIqrc+ijfUps1sLJbkxOl2vsptHb1J0hqA0yX4FDDQ\nfH73ZXcVWdKdREPNfzbcNiOt0n53RNIagMM7mCnEgg1lC8pi1Z/ZuABLj6CJeEw83tlu6+htJNmD\n7YLh4q8ClwNr/YUj+/Y57rKrxOPxuq2rq9Qs/Dd1S18AhbxJ0wgdcfZu61WVylfup+HThYg/i8Lp\nPyRr8BiX1GYe8fqq/91458VXuq2jt0n2EgCYUsA6oDC6c0Nl89aPUy44qHn7OuqWvsDgy25jyDfv\npOHT94hW7j4NQuOahUQrtjD0O/dTOO1qKl68xyW1mYfGmuvEn/UTt3W4QdIbQKSsNA78HTNIyFPz\nwdy3EtHGGpdldYnozk0EhqiWUkMAAAzpSURBVByIx5+NeLxkjRhPZA8fi5SVkjd+CiJC1rCxJJrq\nidVVuKQ4s4g31pZsuG1Grds63CDpDQAgUlb6CfAWMESbI9GGT0pfdFtTVwj0H0XTpg+JN9SQiDbS\nsGYh8ZrdM0zF63biDfXf9dmXX0i81o6I7mni9VUrNt112a1u63ALn9sCusBTwFeArPqVb3yYNWzc\nIb7QgNFui+oM/v4jCBXPoPyx6xF/NoGB+4OkhPemNRqPNkcrt1zgtg43SZlfYaSstBL4JzAYoOa9\np/6tseaUSSGeP+kUhsz6I4MvvhlPdh7+fsN2W+/NK9ytVBCr3Yk3v7C3ZWYU0YrNt2x76CcZnX4u\nZQzA4Q1M8tD+serPautXz3/WbUGdJV5fBUCsppzIx++SO+743dbnhIupW/EqqkrT5lV4soL48vq5\nITUjiNXtXFn15t+ud1uH2yR9N+CeBMPFw4EbgHKgqc9xXz8rMGDUIS7L6pBtD/+UREMteLz0nXI5\nOUWTqV1sks3kHzIdVaXipftoXLsI8TndgEPCLqtOTzQWbW7asurQbY/MyeinP6SgAQAEw8VTgK8D\nayUQ9BeeMvtKT1aufVxaOkXzZ5+WbPnr929wW0cykGpVgBZeBxYDQ7U5Eq1d8vyTqomEy5osKUB0\n56b5VW8/nDGj/ToiJQ0gUlaaAB7A5A/Ma9r04ZbG9ctecVeVJdmJ11eW16+ef56Tc8JCihoA7JpN\n6H7MkGFP7aJn3mnescFmErbslUS0qbFh7eKZVW/+zY4qbUXKGgBApKx0GfAcMAqg6u2Hno3VVax3\nV5Ul2dBEItGwZuEvdsy99XW3tSQbKW0ADk9g2gOGE48lqt566LFEU8TG0Fp20bhh6T/qV7xyu9s6\nkpGUN4BIWWkM+BPwGTAwEalqqC598hGNRRtdlmZJApq2rJ5fu+jZy512I8sepLwBAETKSuuBPwIJ\noCC6fe3OuuUvPm57BjKbpm1ly6vffeycSFlpykSM9jZpYQCwa27B24ECILthzaK1kY/eesrOLpSZ\nNJevLat+9/FzImWlOzreOnNJGwOAXaMG7weGAv76j974MLJ6/tPWBDKL6M6N66sX/POcyOp31rit\nJdlJKwMAiJSVLsDkDxgB+Os/fHVZQ9mCZ60JZAbRyq1bqkufPK9+5Rsr3daSCqSdATi8DPwDYwK+\nuuUvLY58/M6/rQmkN7Hqz7bVvP+vmXXLXsqoxJ5fhpQcC9AZguFiAc4Azgc2ALHccSeMD4499lwR\nT7oaX8bSvH3duurSJ79Vv+LVV93WkkqkrQHALhOYDswENgLRnPCR4byDp8wQry/grjpLd9G46cOV\nNe899YPIxwts6vguktYGALtM4BTgYmAz0BQYHB4YOuKcizyBnAJ31Vm+DKpKw6fvvV+39IXvRcpK\nS93Wk4qkvQG04KQX/xawE6j15vfP7XPMRRd4c/sMd1maZR/QRCJev/K1NyKr518dKSv9yG09qUrG\nGABAMFw8FvgBEAd2iD/L1+fYS8729xs23mVpli6QiDY21C15/rnGDcuuiZSVbnBbTyqTUQYAEAwX\nDwGuwUw/vhkgVDzjhKxhBx0vIq5qs3RMrLp8S3XpE0/Ga3f8KlJWut1tPalOxhkAQDBcHAJmAwdh\neggSOaO/Mjr34BPP8fiz8txVZ9kbqqqN65cur10872ES8fsiZaUpNTdEspKRBgAQDBcHgIuAKcAW\noNGb1y8YKp5xpr/P4LHuqrO0JhFtitQteW5+44ZldwL/cSaLsXQDGWsAsKuH4CjgG0AUk2iUvInT\nDs0Zfdip4vH53dRnaSny//P5eO3O/7GNfd1PRhtAC8Fw8SDg20AYEy8Q8w8oKgwdfvZ53mDBUHfV\nZSYajzVFyt5dXP/h60+D3u/MC2HpZqwBOATDxX5M0NC5QBVQhdfnyZ88vTh7xPgTbOBQ7xGt3Lqu\n5v1/vR+v3fEIMNfJ+WDpAawB7EEwXDwG00DY0ksQ94YG5OUfesZUf7/hk2xPQc+RaG6oqV/5+sKG\nT99/D/hzpKz0U7c1pTvWAPZCMFycC5wJTAMacNoGskdOHJF78JTTvMHQEDf1pRuaiMeaNn+0vHbx\nf1ZotPEJ4PlIWWmz27oyAWsA7eDMQnQxcDDGBOoQkbyJpxyaU3TIFPEFgu4qTG00kYg3l69ZXrf0\nhU/idTvfBx6KlJVucVtXJmENoAOcnoJDgEsx2Ya2AlEJBP15E0/+Svawg462RtA1VBOJ6I4NK2qX\nPr86Xl2+FXgIWGTz9vU+1gA6STBcnIMZVHQWoMA2ICaBoD9vwtQjsoaNPdLjz853VWSSo6oaq9i0\nsnbpi6tilZvLgSeB+ZGyUpvA1SWsAXSRYLi4ENM2cBKtjcDr9+YefOLE7BETjvFk59p5vVuhseaG\n5vK1S+s/enNTrGrrTuAZ4PVIWWnEbW2ZjjWAfcQxglMwkYQenNmKAXL2P2y/7JGTJvv6DhknHq/P\nRZmuEq+v2ty4cfnSyOr5OzTW3ATMA16JlJXWua3NYrAG8CUJhosLgOMwMQTZQDUmjgBPTn5W8IBj\nxmcNOeAQb26fYS7K7DU0HmuKVmxeGfn4nbLmbWURzHfxHFAaKSutdlmeZQ9cMQAR+T6mr/0DVb14\nL+snA0NV9T+tlvmBUlU9VETqVDVPRIYCd6jqjA7O946qHr2X5Q8Ac1X1iS95SS1tBJMwpYL9cIYc\nA40AgUGjB+SMPmKyv++wA9OtiqCx5ki0cuuqpi2rPmlYuyhCPAawHHgR+MjG7icvbhnAKmCqqm5q\nY/0s4HBVvbrVshOB81T1ey0G0A06HqCbDKAFp9dgMFCMaSfIw8QS7MSYAv7CEX2zR04I+wtHhr35\nhUWpWE1INDdURys2r2ratPKTxg3LGtCEF1MFehl4O1JW+pnLEi2doNcNQETuA74JrMZ0/5yDKTo3\nYAblrAU+AXIwkXi/U9XHRORm4A1V/U+rEkAR5gYe75jGuZiuumHAQ6p6g3POlu0FuBM4GRPz3wz8\nRVWfEJHDgNswN+wOYJaqbnVKK1cCMWClql7Q2WsNhot9wFjgeExXogdjAhXO9SL+bF920eT9AoNG\nj/HlFw7zZOcPSkZDSDQ31sRrd6yPVm5e37Rl9bbo9nUAXsx1vAssAj61ATyphVslgHXA4ZgbMKKq\nMRGZCsxW1a+1UQJ4DzhBVSPtGMDvgPFABHgfcxMvbLX9eZiqx6nAIGAlcDnwb+AN4GxV3S4iM4Fp\nqvpNEdkC7KeqTSLSR1Wr9uWag+HiLGB/YAKmdNDXWVWHaTcwxWSP1xMYNGZAoP/Iob4+g4Z4c/sN\n9eT0niloIhHX5oaqRFN9Rby+4rNoxeZNTVs/Lo/XbPcCLfEO1cB8YAmw1hbxUxe3nzQFwIMiEsZ0\nqe11+K2IDAMqVLWjbqOXVHWns89TwLHAwlbrjwMeVdU4sEVEWlJIH4gxjpecWH8vJuAHYBnwsIg8\nDTzdxevbRaSstAn4CPgoGC7+J8aADsCYQdg5p5CIJ5q3rq5t3rp6GWbWYxCP+PsOLfD1GdTXm9uv\njydY0MeTlZvnCeTkSyA7X7z+HMTjFfH48Hh85v3ugxY0EYtqPN5EPNakiWijxmNNGos2JRrrquKR\nqop47Y6KWOXWimjlllo0kQOEHE0AAeBD57tYC2y2QTvpgdsG8CvgNVU913mav97GdqcCL3TieHsW\nZzpbvBHgQ1U9ai/rTscYx5nAL0Rkgqp+qdFpkbLSlviBbcCbwXCxFxgADME0IB7o/DU3oCYkWrEp\nFq3YVI3pbmzG5C9oG4/XI74sr3i9nkRjfTOaaPkuBPN/D2CqWTnOsgSQD+RizG8RUIaphn1mb/j0\nxG0DKMDJywfMarW8FvNjbOFU4PpOHO9kEemHqZeeg2lraM2bwBUi8iAwEDgReATTHjFARI5S1Xed\nHocDME/sEar6moi8DVyAaSPYp2pAWzhF6BZDWAzgmEJ/jDH0wZjDEGdZX0dHy035RaNLxNHmSMuK\nltKAONs2AJXAp5i2kHJMu0QFUG1v9szBbQP4PaYKcB0mSKSF14A5IrIEU68fo6qrOnG89zDhpcMx\njYAL91j/L0zgzkpMLsB3AVS1WURmAHeISAHme7kd+Bh4yFkmmC7Hbr3528Ixhc+c1xdwGhiDjlYf\nprTQ+r0XU0po2uMVdUogFkvyBwKJyLHAJap6ZQfbzWKPhkOLxdI+bpcAOkRV3wbedluHxZKOJH0J\nwGKx9Bx2llyLJYOxBmCxZDDWACyWDMYagMWSwVgDsFgyGGsAFksGYw3AYslgrAFYLBmMNQCLJYOx\nBmCxZDDWACyWDMYagMWSwVgDsFgyGGsAFksGYw3AYslgrAFYLBmMNQCLJYOxBmCxZDDWACyWDMYa\ngMWSwVgDsFgyGGsAFksGYw3AYslg/j+ar0eeJpLH9AAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f9e32a1a278>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 86
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAKkAAAD8CAYAAAAMloRrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWPRxKklPxaKPU0l6KhZ9/H8LUucQk5IlcwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_l78fTflDBs8",
        "colab_type": "text"
      },
      "source": [
        "# Using CRNN\n",
        "This part doesn't work properly. The code for the (attempted) implementation of the CRNN model in PyTorch is here, as mentioned in my report.\n",
        "\n",
        "According to the paper I based it off of, it has seven convolutional layers, forming the CNN part of the network. Then the output of that is fed into a RNN, which is composed of two stacked bi-directional LSTMS. Those are output into a transcription layer, which we utilize a CTC to get the actual string output.\n",
        "\n",
        "The alphabet I defined is below. So is the implementation of the CRNN in Pytorch"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QH5mk4FyjPnS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch.nn as nn\n",
        "\n",
        "\n",
        "alphabet = \"0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ/-\"\n",
        "# Heavily based on the architecture proposed in the paper:\n",
        "# https://arxiv.org/pdf/1507.05717.pdf\n",
        "\n",
        "# \n",
        "class BidirectionalLSTM(nn.Module):\n",
        "\n",
        "    def __init__(self, in_d, hidden_d, out_d):\n",
        "        super(BidirectionalLSTM, self).__init__()\n",
        "        \n",
        "    def forward(self, x):\n",
        "        rec, _ = nn.LSTM(in_d, hidden_d, bidirectional=True)(x)\n",
        "        channels, w, h = rec.size()\n",
        "\n",
        "        out = nn.Linear(2 * hidden_d, out_d)(recurrent.view(channels*w, h))\n",
        "\n",
        "        return out.view(channels, w, -1)\n",
        "\n",
        "\n",
        "class CRNN(nn.Module):\n",
        " # I've added some batchNorms in some places to see\n",
        " # if it would help with the learning\n",
        "    def __init__(self):\n",
        "        super(CRNN, self).__init__()\n",
        "\n",
        "        # 7 Convolution layers as defined in the paper\n",
        "        # for the CNN part of CRNN\n",
        "        # There's a helpful table (Table 1) in there on the 5th page\n",
        "        # which summarizes the architecture of the network\n",
        "        # https://arxiv.org/pdf/1507.05717.pdf\n",
        "        self.conv1 = nn.Sequential(\n",
        "            nn.Conv2d(1, 64, 3, 1, 1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d(2, 2) #stride = 2?\n",
        "        )\n",
        "        self.conv2 = nn.Sequential(\n",
        "            nn.Conv2d(64, 128, 3, 1, 1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d(2, 2)\n",
        "        )\n",
        "        self.conv3 = nn.Sequential(\n",
        "            nn.Conv2d(128, 256, 3, 1, 1),\n",
        "            nn.BatchNorm2d(256),\n",
        "            nn.ReLU(inplace=True)\n",
        "        )\n",
        "        self.conv4 = nn.Sequential(\n",
        "            nn.Conv2d(256, 256, 3, 1, 1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d((2, 2), (2, 1), (0, 1))\n",
        "        )\n",
        "        self.conv5 = nn.Sequential(\n",
        "            nn.Conv2d(256, 512, 3, 1, 1),\n",
        "            nn.BatchNorm2d(512),\n",
        "            nn.ReLU(inplace=True)\n",
        "        )\n",
        "        self.conv6 = nn.Sequential(\n",
        "            nn.Conv2d(512, 512, 3, 1, 1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d((2, 2), (2, 1), (0, 1))\n",
        "        )\n",
        "        self.conv7 = nn.Sequential(\n",
        "            nn.Conv2d(512, 512, 2, 1, 0),\n",
        "            nn.BatchNorm2d(512),\n",
        "            nn.ReLU(inplace=True)\n",
        "        )\n",
        "\n",
        "        # RNN part (recurrent layer)\n",
        "        self.rec = nn.Sequential(\n",
        "            BidirectionalLSTM(512, 256, 256),\n",
        "            BidirectionalLSTM(256, 256, len(alphabet) + 1)\n",
        "        )\n",
        "         # output: len(alphabet) + 1 different 'classes' (the alphabet being used)\n",
        "         # + empty character\n",
        "\n",
        "    def forward(self, x):\n",
        "        # convolutional layer\n",
        "        out = self.conv1(x)\n",
        "        out = self.conv2(out)\n",
        "        out = self.conv3(out)\n",
        "        out = self.conv4(out)\n",
        "        out = self.conv5(out)\n",
        "        out = self.conv6(out)\n",
        "        out = self.conv7(out)\n",
        "\n",
        "        # transform the output of CNN to fit the recurrent layer\n",
        "        out = out.squeeze(2).permute(2, 0, 1)\n",
        "\n",
        "        # recurrent layer\n",
        "        return self.rec(out)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aoKI7aPpq0Q0",
        "colab_type": "code",
        "outputId": "0fcf37d3-4adf-4bec-dca9-1f851466f60b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 33
        }
      },
      "source": [
        "# Alphabet (characters we want to recognize in our nutrition labels)\n",
        "print(len(alphabet) + 1)"
      ],
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "65\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}